(discrete_choice)=

Discrete Choice and Random Utility Models

:::{post} June, 2023 :tags: categorical regression, generalized linear model, discrete choice, model expansion :category: advanced, reference :author: Nathaniel Forde :::

:::{include} ../extra_installs.md :::

import arviz as az
import numpy as np  # For vectorized math operations
import pandas as pd  # For file input/output
import pymc as pm
import pytensor.tensor as pt

from matplotlib import pyplot as plt
from matplotlib.lines import Line2D

%config InlineBackend.figure_format = 'retina'  # high resolution figures
az.style.use("arviz-darkgrid")
rng = np.random.default_rng(42)

Discrete Choice Modelling: The Idea

Discrete choice modelling is related to the idea of a latent utility scale as discussed in {ref}ordinal_regression, but it generalises the idea to decision making. It posits that human decision making can be modelled as a function of latent/subjective utility measurements over a set of mutually exclusive alternatives. The theory states that any decision maker will go with the option that maximises their subjective utility, and that utility can be modelled as a latent linear function of observable features of the world.

The idea is perhaps most famously applied by Daniel McFadden in the 1970s to predict the market share accruing to transportation choices (i.e. car, rail, walking etc..) in California after the proposed introduction of BART light rail system. It's worth pausing on that point. The theory is one of micro level human decision making that has, in real applications, been scaled up to make broadly accurate macro level predictions. For more details we recommend {cite:t}train2009

We don't need to be too credulous either. This is merely a statistical model and success here is entirely dependent on the skill of modeller and the available measurements coupled with plausible theory. But it's worth noting the scale of the ambition underlying these models. The structure of the model encourages you to articulate your theory of the decision makers and the environment they inhabit.

The Data

In this example, we'll examine the technique of discrete choice modelling using a (i) heating system data set from the R mlogit package and (ii) repeat choice data set over cracker brands. We'll be pursuing a Bayesian approach to estimating the models rather than the MLE methodology reported in their vigenette. The first data set shows household choices over offers of heating systems in California. The observations consist of single-family houses in California that were newly built and had central air-conditioning. Five types of systems are considered to have been possible:

• gas central (gc),
• gas room (gr),
• electric central (ec),
• electric room (er),
• heat pump (hp).

The data set reports the installation ic.alt and operating costs oc.alt each household was faced with for each of the five alternatives with some broad demographic information about the household and crucially the choice depvar. This is what one choice scenario over the five alternative looks like in the data:

try:
    wide_heating_df = pd.read_csv("../data/heating_data_r.csv")
except:
    wide_heating_df = pd.read_csv(pm.get_data("heating_data_r.csv"))

wide_heating_df[wide_heating_df["idcase"] == 1]

	idcase	depvar	ic.gc	ic.gr	ic.ec	ic.er	ic.hp	oc.gc	oc.gr	oc.ec	oc.er	oc.hp	income	agehed	rooms	region
0	1	gc	866.0	962.64	859.9	995.76	1135.5	199.69	151.72	553.34	505.6	237.88	7	25	6	ncostl

The core idea of these kinds of models is to conceive of this scenario as a choice over exhaustive options with attached latent utility. The utility ascribed to each option is viewed as a linear combination of the attributes for each option. The utility ascribed to each alternative drives the probability of choosing amongst each option. For each $j$ in all the alternatives $Alt = \{ gc, gr, ec, er, hp \}$ which is assumed to take a Gumbel distribution because this has a particularly nice mathematical property.

$$ \mathbf{U} \sim Gumbel $$$$ \begin{pmatrix} u_{gc} \\ u_{gr} \\ u_{ec} \\ u_{er} \\ u_{hp} \\ \end{pmatrix} = \begin{pmatrix} gc_{ic} & gc_{oc} \\ gr_{ic} & gr_{oc} \\ ec_{ic} & ec_{oc} \\ er_{ic} & er_{oc} \\ hp_{ic} & hp_{oc} \\ \end{pmatrix} \begin{pmatrix} \beta_{ic} \\ \beta_{oc} \\ \end{pmatrix} $$

This assumption proves to be mathematically convenient because the difference between two Gumbel distributions can be modelled as a logistic function, meaning we can model a contrast difference among multiple alternatives with the softmax function. Details of the derivation can be found in {cite:t}train2009

$$ \text{softmax}(u)_{j} = \frac{\exp(u_{j})}{\sum_{q=1}^{J}\exp(u_{q})} $$

The model then assumes that decision maker chooses the option that maximises their subjective utility. The individual utility functions can be richly parameterised. The model is identified just when the utility measures of the alternatives are benchmarked against the fixed utility of the "outside good." The last quantity is fixed at 0.

$$\begin{pmatrix} u_{gc} \\ u_{gr} \\ u_{ec} \\ u_{er} \\ 0 \\ \end{pmatrix} $$

With all these constraints applied we can build out conditional random utility model and it's hierarchical variants. Like nearly all subjects in statistics the precise vocabulary for the model specification is overloaded. The conditional logit parameters $\beta$ may be fixed at the level of the individual, but can vary across individuals and the alternatives gc, gr, ec, er too. In this manner we can compose an elaborate theory of how we expect drivers of subjective utility to change the market share amongst a set of competing goods.

Digression on Data Formats

Discrete choice models are often estimated using a long-data format where each choice scenario is represented with a row per alternative ID and a binary flag denoting the chosen option in each scenario. This data format is recommended for estimating these kinds of models in stan and in pylogit. The reason for doing this is that once the columns installation_costs and operating_costs have been pivoted in this fashion it's easier to include them in matrix calculations.

try:
    long_heating_df = pd.read_csv("../data/long_heating_data.csv")
except:
    long_heating_df = pd.read_csv(pm.get_data("long_heating_data.csv"))

columns = [c for c in long_heating_df.columns if c != "Unnamed: 0"]
long_heating_df[long_heating_df["idcase"] == 1][columns]

	idcase	alt_id	choice	depvar	income	agehed	rooms	region	installation_costs	operating_costs
0	1	1	1	gc	7	25	6	ncostl	866.00	199.69
1	1	2	0	gc	7	25	6	ncostl	962.64	151.72
2	1	3	0	gc	7	25	6	ncostl	859.90	553.34
3	1	4	0	gc	7	25	6	ncostl	995.76	505.60
4	1	5	0	gc	7	25	6	ncostl	1135.50	237.88

The Basic Model

We will show here how to incorporate the utility specifications in PyMC. PyMC is a nice interface for this kind of modelling because it can express the model quite cleanly following the natural mathematical expression for this system of equations. You can see in this simple model how we go about constructing equations for the utility measure of each alternative seperately, and then stacking them together to create the input matrix for our softmax transform.

N = wide_heating_df.shape[0]
observed = pd.Categorical(wide_heating_df["depvar"]).codes
coords = {
    "alts_probs": ["ec", "er", "gc", "gr", "hp"],
    "obs": range(N),
}

with pm.Model(coords=coords) as model_1:
    beta_ic = pm.Normal("beta_ic", 0, 1)
    beta_oc = pm.Normal("beta_oc", 0, 1)

    ## Construct Utility matrix and Pivot
    u0 = beta_ic * wide_heating_df["ic.ec"] + beta_oc * wide_heating_df["oc.ec"]
    u1 = beta_ic * wide_heating_df["ic.er"] + beta_oc * wide_heating_df["oc.er"]
    u2 = beta_ic * wide_heating_df["ic.gc"] + beta_oc * wide_heating_df["oc.gc"]
    u3 = beta_ic * wide_heating_df["ic.gr"] + beta_oc * wide_heating_df["oc.gr"]
    u4 = np.zeros(N)  # Outside Good
    s = pm.math.stack([u0, u1, u2, u3, u4]).T

    ## Apply Softmax Transform
    p_ = pm.Deterministic("p", pm.math.softmax(s, axis=1), dims=("obs", "alts_probs"))

    ## Likelihood
    choice_obs = pm.Categorical("y_cat", p=p_, observed=observed, dims="obs")

    idata_m1 = pm.sample_prior_predictive()
    idata_m1.extend(
        pm.sample(nuts_sampler="numpyro", idata_kwargs={"log_likelihood": True}, random_seed=101)
    )
    idata_m1.extend(pm.sample_posterior_predictive(idata_m1))

pm.model_to_graphviz(model_1)

Sampling: [beta_ic, beta_oc, y_cat]
/Users/nathanielforde/mambaforge/envs/pymc_examples_new/lib/python3.9/site-packages/pymc/sampling/mcmc.py:243: UserWarning: Use of external NUTS sampler is still experimental
  warnings.warn("Use of external NUTS sampler is still experimental", UserWarning)

Compiling...
Compilation time =  0:00:01.199248
Sampling...

  0%|          | 0/2000 [00:00<?, ?it/s]

  0%|          | 0/2000 [00:00<?, ?it/s]

  0%|          | 0/2000 [00:00<?, ?it/s]

  0%|          | 0/2000 [00:00<?, ?it/s]

Sampling time =  0:00:02.165671
Transforming variables...
Transformation time =  0:00:00.404064
Computing Log Likelihood...

idata_m1

arviz.InferenceData

• posterior

  <xarray.Dataset>
  Dimensions:     (chain: 4, draw: 1000, obs: 900, alts_probs: 5)
  Coordinates:
    * chain       (chain) int64 0 1 2 3
    * draw        (draw) int64 0 1 2 3 4 5 6 7 ... 992 993 994 995 996 997 998 999
    * obs         (obs) int64 0 1 2 3 4 5 6 7 ... 892 893 894 895 896 897 898 899
    * alts_probs  (alts_probs) <U2 'ec' 'er' 'gc' 'gr' 'hp'
  Data variables:
      beta_ic     (chain, draw) float64 0.001854 0.001794 ... 0.002087 0.001928
      beta_oc     (chain, draw) float64 -0.003802 -0.003846 ... -0.00447 -0.004533
      p           (chain, draw, obs, alts_probs) float64 0.07322 0.1129 ... 0.1111
  Attributes:
      created_at:     2023-07-13T17:29:12.988497
      arviz_version:  0.15.1

  xarray.Dataset

  • Dimensions:
    • chain: 4
    • draw: 1000
    • obs: 900
    • alts_probs: 5
  • Coordinates: (4)
    • chain
      (chain)
      int64
      0 1 2 3

      array([0, 1, 2, 3])

    • draw
      (draw)
      int64
      0 1 2 3 4 5 ... 995 996 997 998 999

      array([  0,   1,   2, ..., 997, 998, 999])

    • obs
      (obs)
      int64
      0 1 2 3 4 5 ... 895 896 897 898 899

      array([  0,   1,   2, ..., 897, 898, 899])

    • alts_probs
      (alts_probs)
      <U2
      'ec' 'er' 'gc' 'gr' 'hp'

      array(['ec', 'er', 'gc', 'gr', 'hp'], dtype='<U2')

  • Data variables: (3)
    • beta_ic
      (chain, draw)
      float64
      0.001854 0.001794 ... 0.001928

      array([[0.00185432, 0.0017936 , 0.00168995, ..., 0.00179124, 0.0017465 ,
              0.00167666],
             [0.00177376, 0.00165435, 0.00215904, ..., 0.00198785, 0.00193043,
              0.00181715],
             [0.00185793, 0.00232339, 0.00194828, ..., 0.00203171, 0.0021245 ,
              0.00171454],
             [0.0018481 , 0.00185684, 0.00183757, ..., 0.00195163, 0.00208711,
              0.00192813]])

    • beta_oc
      (chain, draw)
      float64
      -0.003802 -0.003846 ... -0.004533

      array([[-0.00380192, -0.00384633, -0.00387485, ..., -0.00462905,
              -0.00462057, -0.00383906],
             [-0.00407646, -0.00419749, -0.00384315, ..., -0.00435413,
              -0.00422983, -0.00419225],
             [-0.00389835, -0.00446452, -0.00464509, ..., -0.00496631,
              -0.00477238, -0.00364794],
             [-0.00398744, -0.00407702, -0.00393378, ..., -0.00416899,
              -0.00447043, -0.00453276]])

    • p
      (chain, draw, obs, alts_probs)
      float64
      0.07322 0.1129 ... 0.4256 0.1111

      array([[[[0.07322309, 0.11294964, 0.2841128 , 0.40787089, 0.12184357],
               [0.09165709, 0.1249422 , 0.30704356, 0.32518664, 0.1511705 ],
               [0.10218969, 0.16264231, 0.23121714, 0.36117826, 0.14277259],
               ...,
               [0.08249413, 0.13373172, 0.23616285, 0.44085863, 0.10675267],
               [0.06839877, 0.17788318, 0.22552951, 0.42617996, 0.10200859],
               [0.08482953, 0.13414628, 0.27236361, 0.40738278, 0.1012778 ]],
      
              [[0.07191539, 0.11025465, 0.28335124, 0.40526019, 0.12921853],
               [0.08962555, 0.12163123, 0.30624134, 0.32372789, 0.15877399],
               [0.10060054, 0.15861184, 0.23209332, 0.35897101, 0.14972329],
               ...,
               [0.08155359, 0.13079867, 0.23744346, 0.43673993, 0.11346434],
               [0.06799272, 0.174111  , 0.2262014 , 0.42287908, 0.1088158 ],
               [0.08351165, 0.1314237 , 0.27347175, 0.40338936, 0.10820353]],
      
              [[0.0707887 , 0.10715561, 0.28156113, 0.39923231, 0.14126225],
               [0.08744061, 0.11758153, 0.30394884, 0.32027513, 0.17075389],
               [0.09898362, 0.15332404, 0.23304048, 0.35392342, 0.16072844],
               ...,
      ...
               ...,
               [0.07434766, 0.12424822, 0.23666514, 0.45907092, 0.10566806],
               [0.06105228, 0.16958401, 0.22476353, 0.4435739 , 0.10102629],
               [0.07642407, 0.12504988, 0.27629164, 0.42177755, 0.10045687]],
      
              [[0.05882476, 0.0966925 , 0.28953088, 0.43895928, 0.11599257],
               [0.07701313, 0.10985028, 0.3211502 , 0.34258706, 0.14939933],
               [0.08840643, 0.15014045, 0.23352096, 0.38783924, 0.14009291],
               ...,
               [0.06805284, 0.11789883, 0.2356187 , 0.4787461 , 0.09968353],
               [0.05517125, 0.164697  , 0.22310792, 0.46191585, 0.095108  ],
               [0.07017311, 0.11890053, 0.2785132 , 0.43781697, 0.0945962 ]],
      
              [[0.05711837, 0.09215532, 0.28711905, 0.42994953, 0.13365773],
               [0.0737155 , 0.10374125, 0.31766931, 0.33720999, 0.16766396],
               [0.08587985, 0.14203293, 0.23520541, 0.38005886, 0.15682296],
               ...,
               [0.06705557, 0.11277456, 0.238708  , 0.46575398, 0.11570788],
               [0.05516607, 0.15734329, 0.22488893, 0.45122393, 0.11137777],
               [0.06842681, 0.11410865, 0.28072918, 0.42561475, 0.11112061]]]])

  • Indexes: (4)
    • chain
      PandasIndex

      PandasIndex(Int64Index([0, 1, 2, 3], dtype='int64', name='chain'))

    • draw
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  990, 991, 992, 993, 994, 995, 996, 997, 998, 999],
                 dtype='int64', name='draw', length=1000))

    • obs
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  890, 891, 892, 893, 894, 895, 896, 897, 898, 899],
                 dtype='int64', name='obs', length=900))

    • alts_probs
      PandasIndex

      PandasIndex(Index(['ec', 'er', 'gc', 'gr', 'hp'], dtype='object', name='alts_probs'))

  • Attributes: (2)

    created_at :

    2023-07-13T17:29:12.988497

    arviz_version :

    0.15.1

  
  
• posterior_predictive

  <xarray.Dataset>
  Dimensions:  (chain: 4, draw: 1000, obs: 900)
  Coordinates:
    * chain    (chain) int64 0 1 2 3
    * draw     (draw) int64 0 1 2 3 4 5 6 7 8 ... 992 993 994 995 996 997 998 999
    * obs      (obs) int64 0 1 2 3 4 5 6 7 8 ... 892 893 894 895 896 897 898 899
  Data variables:
      y_cat    (chain, draw, obs) int64 4 3 4 2 2 1 1 2 4 1 ... 4 3 2 1 4 2 1 1 0
  Attributes:
      created_at:                 2023-07-13T17:29:19.656809
      arviz_version:              0.15.1
      inference_library:          pymc
      inference_library_version:  5.3.0

  xarray.Dataset

  • Dimensions:
    • chain: 4
    • draw: 1000
    • obs: 900
  • Coordinates: (3)
    • chain
      (chain)
      int64
      0 1 2 3

      array([0, 1, 2, 3])

    • draw
      (draw)
      int64
      0 1 2 3 4 5 ... 995 996 997 998 999

      array([  0,   1,   2, ..., 997, 998, 999])

    • obs
      (obs)
      int64
      0 1 2 3 4 5 ... 895 896 897 898 899

      array([  0,   1,   2, ..., 897, 898, 899])

  • Data variables: (1)
    • y_cat
      (chain, draw, obs)
      int64
      4 3 4 2 2 1 1 2 ... 3 2 1 4 2 1 1 0

      array([[[4, 3, 4, ..., 0, 4, 3],
              [3, 1, 3, ..., 3, 3, 1],
              [0, 4, 1, ..., 3, 2, 1],
              ...,
              [3, 3, 4, ..., 3, 2, 2],
              [3, 3, 4, ..., 4, 2, 4],
              [0, 2, 2, ..., 3, 0, 2]],
      
             [[2, 2, 3, ..., 3, 3, 2],
              [4, 0, 1, ..., 3, 4, 1],
              [2, 3, 3, ..., 3, 0, 3],
              ...,
              [3, 2, 1, ..., 2, 2, 3],
              [3, 3, 0, ..., 1, 3, 4],
              [2, 3, 2, ..., 3, 3, 3]],
      
             [[0, 2, 3, ..., 3, 2, 1],
              [3, 2, 3, ..., 3, 3, 3],
              [3, 1, 3, ..., 2, 3, 4],
              ...,
              [2, 4, 4, ..., 3, 1, 3],
              [3, 2, 2, ..., 1, 3, 3],
              [1, 4, 2, ..., 2, 2, 3]],
      
             [[3, 2, 2, ..., 3, 1, 1],
              [3, 2, 2, ..., 2, 3, 3],
              [0, 2, 1, ..., 4, 3, 2],
              ...,
              [3, 2, 3, ..., 3, 1, 3],
              [3, 2, 1, ..., 1, 1, 2],
              [3, 2, 0, ..., 1, 1, 0]]])

  • Indexes: (3)
    • chain
      PandasIndex

      PandasIndex(Int64Index([0, 1, 2, 3], dtype='int64', name='chain'))

    • draw
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  990, 991, 992, 993, 994, 995, 996, 997, 998, 999],
                 dtype='int64', name='draw', length=1000))

    • obs
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  890, 891, 892, 893, 894, 895, 896, 897, 898, 899],
                 dtype='int64', name='obs', length=900))

  • Attributes: (4)

    created_at :

    2023-07-13T17:29:19.656809

    arviz_version :

    0.15.1

    inference_library :

    pymc

    inference_library_version :

    5.3.0

  
  
• log_likelihood

  <xarray.Dataset>
  Dimensions:      (chain: 4, draw: 1000, y_cat_dim_0: 1, y_cat_dim_1: 900)
  Coordinates:
    * chain        (chain) int64 0 1 2 3
    * draw         (draw) int64 0 1 2 3 4 5 6 7 ... 993 994 995 996 997 998 999
    * y_cat_dim_0  (y_cat_dim_0) int64 0
    * y_cat_dim_1  (y_cat_dim_1) int64 0 1 2 3 4 5 6 ... 894 895 896 897 898 899
  Data variables:
      y_cat        (chain, draw, y_cat_dim_0, y_cat_dim_1) float64 -1.258 ... -...
  Attributes:
      created_at:     2023-07-13T17:29:12.992158
      arviz_version:  0.15.1

  xarray.Dataset

  • Dimensions:
    • chain: 4
    • draw: 1000
    • y_cat_dim_0: 1
    • y_cat_dim_1: 900
  • Coordinates: (4)
    • chain
      (chain)
      int64
      0 1 2 3

      array([0, 1, 2, 3])

    • draw
      (draw)
      int64
      0 1 2 3 4 5 ... 995 996 997 998 999

      array([  0,   1,   2, ..., 997, 998, 999])

    • y_cat_dim_0
      (y_cat_dim_0)
      int64
      0

      array([0])

    • y_cat_dim_1
      (y_cat_dim_1)
      int64
      0 1 2 3 4 5 ... 895 896 897 898 899

      array([  0,   1,   2, ..., 897, 898, 899])

  • Data variables: (1)
    • y_cat
      (chain, draw, y_cat_dim_0, y_cat_dim_1)
      float64
      -1.258 -1.181 ... -1.492 -1.27

      array([[[[-1.25838394, -1.18076564, -1.46439801, ..., -1.44323366,
                -1.48930428, -1.30061729]],
      
              [[-1.26106802, -1.18338178, -1.46061574, ..., -1.43782573,
                -1.48632955, -1.29655694]],
      
              [[-1.26740569, -1.19089588, -1.45654309, ..., -1.43093889,
                -1.48261589, -1.29312192]],
      
              ...,
      
              [[-1.25775021, -1.15734211, -1.44256548, ..., -1.4233743 ,
                -1.48837697, -1.26490478]],
      
              [[-1.261385  , -1.16229154, -1.44212494, ..., -1.42140372,
                -1.48738218, -1.26506468]],
      
              [[-1.26872681, -1.19349149, -1.45717302, ..., -1.43107286,
                -1.4822969 , -1.29456549]]],
      
      ...
      
             [[[-1.25609642, -1.17308582, -1.45835841, ..., -1.43792089,
                -1.48815798, -1.29134221]],
      
              [[-1.2546239 , -1.16876045, -1.45611314, ..., -1.43640874,
                -1.48834488, -1.28746152]],
      
              [[-1.25735608, -1.17616661, -1.45956454, ..., -1.43852903,
                -1.48786646, -1.29357521]],
      
              ...,
      
              [[-1.24850717, -1.15764379, -1.45708094, ..., -1.44110905,
                -1.49270641, -1.28629831]],
      
              [[-1.23949333, -1.13584634, -1.45448342, ..., -1.44554047,
                -1.5000997 , -1.27828983]],
      
              [[-1.24785833, -1.14674436, -1.44729605, ..., -1.43251422,
                -1.49214862, -1.27036485]]]])

  • Indexes: (4)
    • chain
      PandasIndex

      PandasIndex(Int64Index([0, 1, 2, 3], dtype='int64', name='chain'))

    • draw
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  990, 991, 992, 993, 994, 995, 996, 997, 998, 999],
                 dtype='int64', name='draw', length=1000))

    • y_cat_dim_0
      PandasIndex

      PandasIndex(Int64Index([0], dtype='int64', name='y_cat_dim_0'))

    • y_cat_dim_1
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  890, 891, 892, 893, 894, 895, 896, 897, 898, 899],
                 dtype='int64', name='y_cat_dim_1', length=900))

  • Attributes: (2)

    created_at :

    2023-07-13T17:29:12.992158

    arviz_version :

    0.15.1

  
  
• sample_stats

  <xarray.Dataset>
  Dimensions:          (chain: 4, draw: 1000)
  Coordinates:
    * chain            (chain) int64 0 1 2 3
    * draw             (draw) int64 0 1 2 3 4 5 6 ... 993 994 995 996 997 998 999
  Data variables:
      acceptance_rate  (chain, draw) float64 0.9667 1.0 0.8868 ... 0.8931 0.9283
      step_size        (chain, draw) float64 0.02965 0.02965 ... 0.03209 0.03209
      diverging        (chain, draw) bool False False False ... False False False
      energy           (chain, draw) float64 1.311e+03 1.311e+03 ... 1.312e+03
      n_steps          (chain, draw) int64 3 3 3 3 1 11 3 7 3 ... 7 7 7 11 7 3 7 3
      tree_depth       (chain, draw) int64 2 2 2 2 1 4 2 3 2 ... 2 3 3 3 4 3 2 3 2
      lp               (chain, draw) float64 1.311e+03 1.311e+03 ... 1.311e+03
  Attributes:
      created_at:     2023-07-13T17:29:12.990891
      arviz_version:  0.15.1

  xarray.Dataset

  • Dimensions:
    • chain: 4
    • draw: 1000
  • Coordinates: (2)
    • chain
      (chain)
      int64
      0 1 2 3

      array([0, 1, 2, 3])

    • draw
      (draw)
      int64
      0 1 2 3 4 5 ... 995 996 997 998 999

      array([  0,   1,   2, ..., 997, 998, 999])

  • Data variables: (7)
    • acceptance_rate
      (chain, draw)
      float64
      0.9667 1.0 0.8868 ... 0.8931 0.9283

      array([[0.96668742, 1.        , 0.88678803, ..., 0.832775  , 0.80147234,
              0.97320939],
             [0.91898504, 0.73812295, 0.65193552, ..., 1.        , 0.99555924,
              0.95500129],
             [0.97029319, 0.5078157 , 1.        , ..., 0.97650608, 1.        ,
              1.        ],
             [1.        , 0.98428559, 0.94659264, ..., 0.87880027, 0.89313587,
              0.92827725]])

    • step_size
      (chain, draw)
      float64
      0.02965 0.02965 ... 0.03209 0.03209

      array([[0.02965364, 0.02965364, 0.02965364, ..., 0.02965364, 0.02965364,
              0.02965364],
             [0.0348108 , 0.0348108 , 0.0348108 , ..., 0.0348108 , 0.0348108 ,
              0.0348108 ],
             [0.0335617 , 0.0335617 , 0.0335617 , ..., 0.0335617 , 0.0335617 ,
              0.0335617 ],
             [0.03209221, 0.03209221, 0.03209221, ..., 0.03209221, 0.03209221,
              0.03209221]])

    • diverging
      (chain, draw)
      bool
      False False False ... False False

      array([[False, False, False, ..., False, False, False],
             [False, False, False, ..., False, False, False],
             [False, False, False, ..., False, False, False],
             [False, False, False, ..., False, False, False]])

    • energy
      (chain, draw)
      float64
      1.311e+03 1.311e+03 ... 1.312e+03

      array([[1311.02807455, 1310.85624244, 1311.29145993, ..., 1313.34508594,
              1314.36026935, 1314.50846674],
             [1310.3936618 , 1312.79557002, 1315.79936109, ..., 1310.19152711,
              1310.69088735, 1310.18922683],
             [1310.66343379, 1316.9728028 , 1314.59817151, ..., 1313.57708469,
              1313.3000676 , 1311.96972255],
             [1310.81727096, 1310.15989558, 1310.50270126, ..., 1312.29193502,
              1312.1408833 , 1311.77428022]])

    • n_steps
      (chain, draw)
      int64
      3 3 3 3 1 11 3 7 ... 7 7 11 7 3 7 3

      array([[ 3,  3,  3, ...,  3,  1,  9],
             [ 3,  3,  3, ...,  3,  7,  5],
             [ 7, 11,  3, ...,  3,  3, 15],
             [ 3,  3,  3, ...,  3,  7,  3]])

    • tree_depth
      (chain, draw)
      int64
      2 2 2 2 1 4 2 3 ... 3 3 3 4 3 2 3 2

      array([[2, 2, 2, ..., 2, 1, 4],
             [2, 2, 2, ..., 2, 3, 3],
             [3, 4, 2, ..., 2, 2, 4],
             [2, 2, 2, ..., 2, 3, 2]])

    • lp
      (chain, draw)
      float64
      1.311e+03 1.311e+03 ... 1.311e+03

      array([[1310.85296473, 1310.5090792 , 1311.05196078, ..., 1313.07351323,
              1313.96754755, 1311.23235692],
             [1310.26682798, 1312.48985299, 1315.52133628, ..., 1310.07833465,
              1309.85428101, 1310.09987862],
             [1310.37326154, 1315.2401948 , 1311.20326201, ..., 1313.38619846,
              1311.99334014, 1311.60532063],
             [1310.05439245, 1309.89541318, 1310.19960031, ..., 1309.97580287,
              1310.84659945, 1310.64836248]])

  • Indexes: (2)
    • chain
      PandasIndex

      PandasIndex(Int64Index([0, 1, 2, 3], dtype='int64', name='chain'))

    • draw
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  990, 991, 992, 993, 994, 995, 996, 997, 998, 999],
                 dtype='int64', name='draw', length=1000))

  • Attributes: (2)

    created_at :

    2023-07-13T17:29:12.990891

    arviz_version :

    0.15.1

  
  
• prior

  <xarray.Dataset>
  Dimensions:     (chain: 1, draw: 500, obs: 900, alts_probs: 5)
  Coordinates:
    * chain       (chain) int64 0
    * draw        (draw) int64 0 1 2 3 4 5 6 7 ... 492 493 494 495 496 497 498 499
    * obs         (obs) int64 0 1 2 3 4 5 6 7 ... 892 893 894 895 896 897 898 899
    * alts_probs  (alts_probs) <U2 'ec' 'er' 'gc' 'gr' 'hp'
  Data variables:
      beta_oc     (chain, draw) float64 0.05221 0.1678 0.7865 ... 0.5232 -0.09242
      p           (chain, draw, obs, alts_probs) float64 4.778e-201 ... 4.831e-91
      beta_ic     (chain, draw) float64 -0.57 -0.8283 0.2129 ... 1.097 0.2006
  Attributes:
      created_at:                 2023-07-13T17:29:07.842667
      arviz_version:              0.15.1
      inference_library:          pymc
      inference_library_version:  5.3.0

  xarray.Dataset

  • Dimensions:
    • chain: 1
    • draw: 500
    • obs: 900
    • alts_probs: 5
  • Coordinates: (4)
    • chain
      (chain)
      int64
      0

      array([0])

    • draw
      (draw)
      int64
      0 1 2 3 4 5 ... 495 496 497 498 499

      array([  0,   1,   2, ..., 497, 498, 499])

    • obs
      (obs)
      int64
      0 1 2 3 4 5 ... 895 896 897 898 899

      array([  0,   1,   2, ..., 897, 898, 899])

    • alts_probs
      (alts_probs)
      <U2
      'ec' 'er' 'gc' 'gr' 'hp'

      array(['ec', 'er', 'gc', 'gr', 'hp'], dtype='<U2')

  • Data variables: (3)
    • beta_oc
      (chain, draw)
      float64
      0.05221 0.1678 ... 0.5232 -0.09242

      array([[ 5.22123131e-02,  1.67808803e-01,  7.86499718e-01,
               6.78099653e-01, -9.53864839e-02,  5.76929240e-01,
               2.66210247e-01,  2.76490816e-01, -8.62046317e-01,
               1.43800720e-01,  6.05778337e-01, -5.01770894e-01,
              -1.01903815e+00,  1.44285382e+00,  1.44191513e+00,
              -1.72887211e+00,  2.58057973e-02, -1.40841177e+00,
               1.49409275e-01, -1.25941954e+00,  6.03880072e-01,
               6.44329176e-03,  1.55331894e+00,  1.19140854e+00,
               1.19125366e+00,  8.95483429e-02, -4.88048425e-01,
               2.38948238e-01,  6.13491798e-02, -1.05372724e-01,
               6.14166519e-01, -7.98300206e-01, -1.57627791e+00,
              -6.56281227e-01, -3.31887590e-01, -9.76469578e-02,
               3.73667319e-02, -4.46401625e-01, -6.32421376e-01,
              -5.15888903e-02,  4.91398108e-01,  2.65630805e-01,
              -3.36713237e-01, -6.06621909e-01,  4.64579807e-01,
               1.22095804e+00,  1.41716046e+00,  1.49022768e+00,
              -1.46466745e-02,  3.49249386e-01, -2.01548779e+00,
               1.37776656e+00,  5.88994694e-01,  1.13044210e+00,
              -4.99040316e-02, -1.07193117e+00, -9.94516858e-01,
              -1.26196021e+00, -1.28986812e+00, -1.91050443e+00,
      ...
               2.19951294e+00, -1.22471025e+00,  8.12392922e-01,
              -8.76963801e-01, -6.68999739e-01, -3.79367158e-01,
               3.21759821e-01, -8.71934942e-01, -7.71975654e-01,
               2.20439637e-01, -1.40235273e-02,  6.37145014e-02,
              -7.07404789e-02,  6.94661764e-01, -7.08752140e-01,
              -7.49656515e-01,  7.63016827e-02, -1.96157335e+00,
               4.29880784e-01,  2.26642863e-01, -1.33725618e+00,
              -1.33565769e+00, -6.20769298e-02, -1.56438832e+00,
               1.01849028e+00, -4.47346052e-01, -9.22840328e-01,
               1.52578096e+00, -1.01157454e+00,  5.01524450e-01,
              -1.80982817e-02, -2.56621616e-01, -1.49459668e-01,
               4.00916871e-03,  1.72359937e+00, -9.75697269e-01,
               1.15033052e+00, -7.77945279e-01,  6.28044749e-01,
               1.94636242e-01, -8.22248275e-01, -3.02361708e-01,
              -5.24404285e-01, -8.16118103e-01,  3.66017430e-01,
               3.54466167e-01,  1.39027437e+00, -3.03677763e-02,
              -1.36836882e+00,  1.59254769e+00, -1.10448020e+00,
               1.51725654e+00,  1.32489420e+00,  5.48989024e-01,
              -9.51623984e-01,  1.55167656e+00, -5.33382582e-01,
               5.23223875e-01, -9.24236936e-02]])

    • p
      (chain, draw, obs, alts_probs)
      float64
      4.778e-201 9.217e-236 ... 4.831e-91

      array([[[[4.77797773e-201, 9.21735126e-236, 1.41251546e-210,
                1.37733934e-235, 1.00000000e+000],
               [3.50043885e-186, 3.55737733e-211, 4.22245926e-177,
                9.15008836e-185, 1.00000000e+000],
               [5.68528693e-169, 2.26822738e-214, 2.25678364e-145,
                1.90922389e-191, 1.00000000e+000],
               ...,
               [1.68247109e-198, 2.50560604e-248, 1.32722173e-191,
                9.93208296e-259, 1.00000000e+000],
               [1.77227549e-187, 1.14210303e-269, 5.67585094e-209,
                5.69496368e-265, 1.00000000e+000],
               [1.45508677e-228, 1.40393393e-260, 4.91739513e-218,
                7.13059332e-274, 1.00000000e+000]],
      
              [[1.02747845e-269, 4.59481051e-322, 1.10683126e-297,
                0.00000000e+000, 1.00000000e+000],
               [1.94990948e-249, 4.22213523e-287, 2.80445009e-250,
                2.04748449e-261, 1.00000000e+000],
               [1.14035008e-227, 5.22505369e-295, 2.68012112e-204,
                2.35284962e-272, 1.00000000e+000],
      ...
               [1.84685777e-088, 1.00000000e+000, 4.89269192e-184,
                7.38762399e-064, 0.00000000e+000],
               [1.35107715e-133, 1.00000000e+000, 3.44031578e-169,
                1.46510116e-077, 0.00000000e+000],
               [4.53965711e-046, 1.00000000e+000, 3.19458160e-159,
                1.49103047e-050, 0.00000000e+000]],
      
              [[8.52094383e-026, 4.79059477e-012, 4.53943617e-011,
                1.00000000e+000, 1.74643648e-078],
               [1.38847319e-011, 1.05170246e-001, 1.79553539e-003,
                8.93034219e-001, 4.16854810e-060],
               [1.94441605e-018, 2.32528695e-001, 6.03540293e-018,
                7.67471305e-001, 1.62594133e-063],
               ...,
               [9.17447962e-033, 2.93347875e-014, 2.02552628e-025,
                1.00000000e+000, 5.95661433e-087],
               [4.71956594e-039, 4.02493334e-007, 2.37477741e-022,
                9.99999598e-001, 7.76581052e-089],
               [1.47529013e-027, 2.33452643e-014, 3.10069646e-020,
                1.00000000e+000, 4.83126180e-091]]]])

    • beta_ic
      (chain, draw)
      float64
      -0.57 -0.8283 ... 1.097 0.2006

      array([[-5.70004358e-01, -8.28263292e-01,  2.12865719e-01,
              -1.22589604e+00, -7.38455183e-01,  1.38562947e+00,
              -5.14147314e-01,  8.78378147e-01,  6.29759498e-01,
              -4.48075605e-01, -1.65703227e-01,  9.40024262e-01,
              -1.30760455e-02, -4.49690187e-01,  6.11037300e-01,
              -9.42668616e-01,  1.70679147e+00, -1.52535825e+00,
               1.57013610e+00,  1.83215761e-01,  2.38784640e-04,
              -4.95682078e-01, -4.66365955e-01,  2.51903440e-01,
               3.30163556e-01, -4.77494859e-03, -2.09518470e+00,
               6.23561933e-01, -1.23929521e+00, -2.32219278e-01,
              -4.25066724e-01,  1.18733217e-01,  3.72339384e-01,
              -1.30904663e-01, -1.40000993e-01, -1.58417383e+00,
              -4.68317692e-01, -7.11586976e-01,  2.43477244e-01,
               9.58822534e-01, -5.11833848e-03,  3.42326203e-01,
              -2.29935048e-01,  1.59822529e+00,  1.33733438e+00,
              -5.59368283e-01, -4.92588819e-01, -5.61527404e-01,
              -1.87151514e+00, -5.51304365e-01,  7.80311840e-01,
               1.46567444e+00,  1.52603877e-01,  2.36093837e+00,
              -6.71137980e-01, -4.61363625e-01, -1.32707128e+00,
              -7.97252508e-02, -4.19876746e-01, -4.82346713e-01,
      ...
              -1.61813217e+00,  3.23956615e-01, -2.58347090e-01,
               1.09891379e+00,  1.04963130e-02,  1.21951803e+00,
              -8.88469247e-01,  2.40129297e-01,  8.74456788e-02,
              -1.14112180e-01,  9.70872675e-01,  5.79073453e-02,
               1.95989732e-01,  1.34320229e+00,  8.83436743e-03,
               2.61360349e-01,  1.89287929e-01,  1.42289282e+00,
              -4.10941567e-01, -3.60285728e-02,  6.54785228e-01,
              -1.33499268e-02, -8.97305571e-02,  8.75881939e-01,
               1.49457334e+00, -3.00572733e-01, -9.10286901e-01,
               1.08817750e+00, -1.08538576e+00,  8.05863335e-01,
              -9.09648592e-01, -4.31780695e-01,  7.69342341e-01,
              -9.71707445e-01, -1.54695370e+00,  9.20746449e-01,
               1.12597739e+00,  1.21455611e+00,  8.84523243e-01,
              -1.86645209e+00,  6.68284401e-01, -1.05267736e-01,
               2.12913417e+00, -3.54292999e-01,  2.77048772e-02,
               2.17053506e+00, -9.55429129e-01, -7.62062828e-01,
               2.31270806e+00, -5.42923830e-01, -1.20304858e+00,
              -4.31171792e-01, -8.12882960e-01,  9.24038908e-01,
              -1.41725247e-01,  2.80235277e-01, -6.97680233e-01,
               1.09738313e+00,  2.00559485e-01]])

  • Indexes: (4)
    • chain
      PandasIndex

      PandasIndex(Int64Index([0], dtype='int64', name='chain'))

    • draw
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  490, 491, 492, 493, 494, 495, 496, 497, 498, 499],
                 dtype='int64', name='draw', length=500))

    • obs
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  890, 891, 892, 893, 894, 895, 896, 897, 898, 899],
                 dtype='int64', name='obs', length=900))

    • alts_probs
      PandasIndex

      PandasIndex(Index(['ec', 'er', 'gc', 'gr', 'hp'], dtype='object', name='alts_probs'))

  • Attributes: (4)

    created_at :

    2023-07-13T17:29:07.842667

    arviz_version :

    0.15.1

    inference_library :

    pymc

    inference_library_version :

    5.3.0

  
  
• prior_predictive

  <xarray.Dataset>
  Dimensions:  (chain: 1, draw: 500, obs: 900)
  Coordinates:
    * chain    (chain) int64 0
    * draw     (draw) int64 0 1 2 3 4 5 6 7 8 ... 492 493 494 495 496 497 498 499
    * obs      (obs) int64 0 1 2 3 4 5 6 7 8 ... 892 893 894 895 896 897 898 899
  Data variables:
      y_cat    (chain, draw, obs) int64 4 4 4 4 4 4 4 4 4 4 ... 3 1 1 3 3 3 3 3 3
  Attributes:
      created_at:                 2023-07-13T17:29:07.844409
      arviz_version:              0.15.1
      inference_library:          pymc
      inference_library_version:  5.3.0

  xarray.Dataset

  • Dimensions:
    • chain: 1
    • draw: 500
    • obs: 900
  • Coordinates: (3)
    • chain
      (chain)
      int64
      0

      array([0])

    • draw
      (draw)
      int64
      0 1 2 3 4 5 ... 495 496 497 498 499

      array([  0,   1,   2, ..., 497, 498, 499])

    • obs
      (obs)
      int64
      0 1 2 3 4 5 ... 895 896 897 898 899

      array([  0,   1,   2, ..., 897, 898, 899])

  • Data variables: (1)
    • y_cat
      (chain, draw, obs)
      int64
      4 4 4 4 4 4 4 4 ... 1 1 3 3 3 3 3 3

      array([[[4, 4, 4, ..., 4, 4, 4],
              [4, 4, 4, ..., 4, 4, 4],
              [0, 0, 1, ..., 1, 0, 0],
              ...,
              [4, 4, 4, ..., 4, 4, 4],
              [1, 1, 1, ..., 1, 1, 1],
              [3, 3, 3, ..., 3, 3, 3]]])

  • Indexes: (3)
    • chain
      PandasIndex

      PandasIndex(Int64Index([0], dtype='int64', name='chain'))

    • draw
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  490, 491, 492, 493, 494, 495, 496, 497, 498, 499],
                 dtype='int64', name='draw', length=500))

    • obs
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  890, 891, 892, 893, 894, 895, 896, 897, 898, 899],
                 dtype='int64', name='obs', length=900))

  • Attributes: (4)

    created_at :

    2023-07-13T17:29:07.844409

    arviz_version :

    0.15.1

    inference_library :

    pymc

    inference_library_version :

    5.3.0

  
  
• observed_data

  <xarray.Dataset>
  Dimensions:  (obs: 900)
  Coordinates:
    * obs      (obs) int64 0 1 2 3 4 5 6 7 8 ... 892 893 894 895 896 897 898 899
  Data variables:
      y_cat    (obs) int64 2 2 2 1 1 2 2 2 2 2 2 2 2 ... 2 2 2 2 2 2 2 1 2 2 2 2 2
  Attributes:
      created_at:                 2023-07-13T17:29:07.844758
      arviz_version:              0.15.1
      inference_library:          pymc
      inference_library_version:  5.3.0

  xarray.Dataset

  • Dimensions:
    • obs: 900
  • Coordinates: (1)
    • obs
      (obs)
      int64
      0 1 2 3 4 5 ... 895 896 897 898 899

      array([  0,   1,   2, ..., 897, 898, 899])

  • Data variables: (1)
    • y_cat
      (obs)
      int64
      2 2 2 1 1 2 2 2 ... 2 2 1 2 2 2 2 2

      array([2, 2, 2, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 3, 2, 3, 4, 4, 1, 2, 0, 3,
             2, 4, 2, 3, 2, 0, 0, 2, 2, 3, 3, 2, 2, 2, 3, 2, 2, 1, 2, 3, 2, 2,
             2, 2, 1, 2, 2, 2, 2, 2, 2, 2, 4, 3, 2, 4, 2, 2, 2, 2, 2, 2, 3, 2,
             2, 2, 3, 2, 3, 2, 1, 4, 1, 2, 2, 2, 1, 3, 2, 2, 2, 4, 3, 2, 2, 2,
             2, 2, 4, 2, 2, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 0, 2, 2, 2, 4, 2,
             2, 2, 4, 1, 0, 2, 2, 2, 2, 2, 3, 1, 2, 2, 2, 2, 2, 2, 2, 3, 2, 3,
             0, 3, 2, 3, 1, 0, 2, 2, 3, 2, 1, 4, 4, 3, 2, 3, 2, 2, 4, 3, 4, 2,
             2, 2, 2, 2, 2, 0, 2, 2, 2, 0, 2, 0, 2, 2, 2, 0, 1, 2, 2, 3, 2, 2,
             2, 2, 3, 3, 1, 2, 2, 2, 2, 2, 2, 3, 2, 2, 2, 4, 2, 2, 2, 2, 2, 2,
             2, 2, 0, 2, 2, 3, 2, 2, 2, 2, 2, 2, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2,
             2, 0, 2, 2, 2, 2, 2, 2, 2, 1, 2, 2, 2, 2, 2, 3, 2, 0, 2, 2, 2, 2,
             2, 3, 2, 2, 3, 2, 1, 0, 0, 4, 2, 2, 2, 2, 2, 1, 2, 3, 3, 3, 2, 2,
             2, 3, 2, 2, 2, 2, 2, 2, 3, 2, 2, 2, 2, 2, 3, 2, 2, 2, 4, 0, 2, 1,
             3, 1, 2, 3, 2, 3, 0, 0, 4, 2, 2, 2, 2, 2, 0, 1, 2, 2, 2, 2, 2, 2,
             2, 1, 1, 2, 1, 3, 1, 2, 4, 2, 2, 2, 2, 2, 4, 2, 2, 2, 2, 3, 2, 2,
             2, 2, 1, 3, 2, 3, 2, 2, 2, 4, 4, 2, 3, 3, 3, 1, 2, 2, 2, 2, 2, 2,
             4, 2, 2, 2, 4, 2, 0, 2, 3, 4, 2, 2, 2, 2, 2, 2, 2, 4, 0, 2, 1, 3,
             0, 1, 2, 2, 2, 2, 2, 1, 2, 4, 3, 2, 2, 2, 2, 3, 2, 2, 2, 2, 2, 2,
             2, 0, 2, 2, 2, 2, 2, 2, 1, 2, 2, 1, 0, 3, 0, 2, 2, 3, 3, 2, 1, 0,
             2, 1, 2, 2, 2, 2, 3, 0, 2, 2, 2, 2, 2, 3, 2, 2, 2, 2, 2, 2, 2, 0,
      ...
             3, 2, 1, 2, 3, 2, 3, 3, 3, 2, 0, 2, 2, 3, 1, 2, 0, 1, 3, 2, 1, 0,
             2, 3, 0, 0, 1, 3, 1, 2, 2, 2, 2, 4, 2, 1, 2, 2, 2, 2, 1, 2, 2, 3,
             3, 2, 2, 2, 1, 2, 4, 2, 3, 0, 2, 1, 2, 1, 2, 2, 3, 2, 2, 2, 2, 3,
             2, 2, 2, 2, 2, 1, 3, 2, 3, 2, 2, 2, 2, 3, 2, 2, 2, 1, 2, 0, 2, 2,
             2, 3, 2, 3, 2, 0, 2, 3, 3, 2, 2, 3, 2, 2, 2, 2, 2, 2, 2, 4, 3, 2,
             3, 2, 2, 2, 2, 0, 2, 2, 3, 0, 2, 1, 1, 3, 2, 2, 2, 2, 3, 0, 3, 2,
             2, 2, 2, 2, 2, 2, 2, 2, 4, 0, 4, 2, 1, 2, 1, 2, 2, 0, 2, 2, 1, 2,
             3, 2, 3, 3, 2, 2, 4, 1, 2, 2, 2, 2, 4, 4, 1, 2, 2, 2, 1, 2, 1, 3,
             3, 0, 4, 2, 2, 2, 2, 2, 2, 2, 2, 3, 2, 2, 2, 2, 3, 4, 2, 1, 1, 0,
             2, 2, 2, 2, 1, 1, 3, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 3, 0, 2, 2, 2,
             1, 2, 3, 4, 2, 1, 2, 3, 2, 3, 2, 2, 2, 1, 2, 1, 2, 2, 2, 2, 2, 2,
             2, 2, 2, 4, 2, 2, 2, 2, 2, 1, 2, 0, 2, 2, 3, 0, 3, 2, 2, 2, 3, 3,
             2, 3, 2, 2, 2, 2, 2, 1, 4, 3, 2, 2, 4, 0, 3, 2, 1, 1, 2, 2, 2, 2,
             0, 2, 2, 0, 2, 3, 2, 2, 2, 2, 3, 3, 2, 2, 2, 3, 2, 2, 2, 1, 4, 0,
             2, 2, 2, 2, 2, 2, 2, 1, 4, 2, 2, 0, 2, 1, 2, 2, 0, 2, 3, 2, 2, 3,
             3, 4, 2, 3, 2, 0, 2, 3, 2, 4, 3, 3, 2, 2, 2, 1, 2, 2, 1, 2, 2, 3,
             0, 2, 2, 3, 2, 2, 1, 2, 3, 2, 2, 2, 2, 4, 1, 2, 2, 2, 2, 2, 3, 2,
             1, 2, 1, 0, 2, 2, 2, 2, 2, 2, 2, 3, 2, 3, 0, 2, 1, 2, 2, 2, 2, 2,
             1, 1, 2, 2, 2, 2, 2, 0, 2, 2, 2, 2, 2, 1, 3, 2, 2, 4, 4, 2, 2, 2,
             1, 2, 2, 2, 2, 2, 0, 2, 2, 2, 2, 2, 2, 2, 1, 2, 2, 2, 2, 2])

  • Indexes: (1)
    • obs
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  890, 891, 892, 893, 894, 895, 896, 897, 898, 899],
                 dtype='int64', name='obs', length=900))

  • Attributes: (4)

    created_at :

    2023-07-13T17:29:07.844758

    arviz_version :

    0.15.1

    inference_library :

    pymc

    inference_library_version :

    5.3.0

  
  

summaries = az.summary(idata_m1, var_names=["beta_ic", "beta_oc"])
summaries

	mean	sd	hdi_3%	hdi_97%	mcse_mean	mcse_sd	ess_bulk	ess_tail	r_hat
beta_ic	0.002	0.0	0.002	0.002	0.0	0.0	1870.0	1999.0	1.0
beta_oc	-0.004	0.0	-0.005	-0.004	0.0	0.0	881.0	1105.0	1.0

In the mlogit vignette they report how the above model specification leads to inadequate parameter estimates. They note for instance that while the utility scale itself is hard to interpret the value of the ratio of the coefficients is often meaningful because when:

$$ U = \beta_{oc}oc + \beta_{ic}ic $$

then the marginal rate of substitution is just the ratio of the two beta coefficients. The relative importance of one component of the utility equation to another is an economically meaningful quantity even if the notion of subjective utility is itself unobservable.

$$ dU = \beta_{ic} dic + \beta_{oc} doc = 0 \Rightarrow -\frac{dic}{doc}\mid_{dU=0}=\frac{\beta_{oc}}{\beta_{ic}}$$

Our parameter estimates differ slightly from the reported estimates, but we agree the model is inadequate. We will show a number of Bayesian model checks to demonstrate this fact, but the main call out is that the parameter values for installation costs should probably be negative. It's counter-intuitive that a $\beta_{ic}$ increase in price would increase the utility of generated by the installation even marginally as here. Although we might imagine that some kind of quality assurance comes with price which drives satisfaction with higher installation costs. The coefficient for repeat operating costs is negative as expected. Putting this issue aside for now, we'll show below how we can incorporate prior knowledge to adjust for this kind of conflicts with theory.

But in any case, once we have a fitted model we can calculate the marginal rate of substitution as follows:

## marginal rate of substitution for a reduction in installation costs
post = az.extract(idata_m1)
substitution_rate = post["beta_oc"] / post["beta_ic"]
substitution_rate.mean().item()

-2.215565260325244

This statistic gives a view of the relative importance of the attributes which drive our utility measures. But being good Bayesians we actually want to calculate the posterior distribution for this statistic.

fig, ax = plt.subplots(figsize=(20, 10))

ax.hist(
    substitution_rate,
    bins=30,
    ec="black",
)
ax.set_title("Uncertainty in Marginal Rate of Substitution \n Operating Costs / Installation Costs");

which suggests that there is almost twice the value accorded to the a unit reduction in recurring operating costs over the one-off installation costs. Whether this is remotely plausible is almost beside the point since the model does not even closely capture the data generating process. But it's worth repeating that the native scale of utility is not straightforwardly meaningful, but the ratio of the coefficients in the utility equations can be directly interpreted.

To assess overall model adequacy we can rely on the posterior predictive checks to see if the model can recover an approximation to the data generating process.

idata_m1["posterior"]["p"].mean(dim=["chain", "draw", "obs"])

<xarray.DataArray 'p' (alts_probs: 5)>
array([0.08414602, 0.13748865, 0.26918857, 0.38213088, 0.12704589])
Coordinates:
  * alts_probs  (alts_probs) <U2 'ec' 'er' 'gc' 'gr' 'hp'

xarray.DataArray

'p'

• alts_probs: 5

• 0.08415 0.1375 0.2692 0.3821 0.127

  array([0.08414602, 0.13748865, 0.26918857, 0.38213088, 0.12704589])

• Coordinates: (1)
  • alts_probs
    (alts_probs)
    <U2
    'ec' 'er' 'gc' 'gr' 'hp'

    array(['ec', 'er', 'gc', 'gr', 'hp'], dtype='<U2')

• Indexes: (1)
  • alts_probs
    PandasIndex

    PandasIndex(Index(['ec', 'er', 'gc', 'gr', 'hp'], dtype='object', name='alts_probs'))

• Attributes: (0)

fig, axs = plt.subplots(1, 2, figsize=(20, 10))
ax = axs[0]
counts = wide_heating_df.groupby("depvar")["idcase"].count()
predicted_shares = idata_m1["posterior"]["p"].mean(dim=["chain", "draw", "obs"])
ci_lb = idata_m1["posterior"]["p"].quantile(0.025, dim=["chain", "draw", "obs"])
ci_ub = idata_m1["posterior"]["p"].quantile(0.975, dim=["chain", "draw", "obs"])
ax.scatter(ci_lb, ["ec", "er", "gc", "gr", "hp"], color="k", s=2)
ax.scatter(ci_ub, ["ec", "er", "gc", "gr", "hp"], color="k", s=2)
ax.scatter(
    counts / counts.sum(),
    ["ec", "er", "gc", "gr", "hp"],
    label="Observed Shares",
    color="red",
    s=100,
)
ax.hlines(
    ["ec", "er", "gc", "gr", "hp"], ci_lb, ci_ub, label="Predicted 95% Interval", color="black"
)
ax.legend()
ax.set_title("Observed V Predicted Shares")
az.plot_ppc(idata_m1, ax=axs[1])
axs[1].set_title("Posterior Predictive Checks")
ax.set_xlabel("Shares")
ax.set_ylabel("Heating System");

We can see here that the model is fairly inadequate, and fails quite dramatically to recapture the posterior predictive distribution.

Improved Model: Adding Alternative Specific Intercepts

We can address some of the issues with the prior model specification by adding intercept terms for each of the unique alternatives gr, gc, ec, er. These terms will absorb some of the error seen in the last model by allowing us to control some of the heterogenity of utility measures across products.

N = wide_heating_df.shape[0]
observed = pd.Categorical(wide_heating_df["depvar"]).codes

coords = {
    "alts_intercepts": ["ec", "er", "gc", "gr"],
    "alts_probs": ["ec", "er", "gc", "gr", "hp"],
    "obs": range(N),
}
with pm.Model(coords=coords) as model_2:
    beta_ic = pm.Normal("beta_ic", 0, 1)
    beta_oc = pm.Normal("beta_oc", 0, 1)
    alphas = pm.Normal("alpha", 0, 1, dims="alts_intercepts")

    ## Construct Utility matrix and Pivot using an intercept per alternative
    u0 = alphas[0] + beta_ic * wide_heating_df["ic.ec"] + beta_oc * wide_heating_df["oc.ec"]
    u1 = alphas[1] + beta_ic * wide_heating_df["ic.er"] + beta_oc * wide_heating_df["oc.er"]
    u2 = alphas[2] + beta_ic * wide_heating_df["ic.gc"] + beta_oc * wide_heating_df["oc.gc"]
    u3 = alphas[3] + beta_ic * wide_heating_df["ic.gr"] + beta_oc * wide_heating_df["oc.gr"]
    u4 = np.zeros(N)  # Outside Good
    s = pm.math.stack([u0, u1, u2, u3, u4]).T

    ## Apply Softmax Transform
    p_ = pm.Deterministic("p", pm.math.softmax(s, axis=1), dims=("obs", "alts_probs"))

    ## Likelihood
    choice_obs = pm.Categorical("y_cat", p=p_, observed=observed, dims="obs")

    idata_m2 = pm.sample_prior_predictive()
    idata_m2.extend(
        pm.sample(nuts_sampler="numpyro", idata_kwargs={"log_likelihood": True}, random_seed=103)
    )
    idata_m2.extend(pm.sample_posterior_predictive(idata_m2))


pm.model_to_graphviz(model_2)

Sampling: [alpha, beta_ic, beta_oc, y_cat]
/Users/nathanielforde/mambaforge/envs/pymc_examples_new/lib/python3.9/site-packages/pymc/sampling/mcmc.py:243: UserWarning: Use of external NUTS sampler is still experimental
  warnings.warn("Use of external NUTS sampler is still experimental", UserWarning)

Compiling...
Compilation time =  0:00:01.339706
Sampling...

  0%|          | 0/2000 [00:00<?, ?it/s]

  0%|          | 0/2000 [00:00<?, ?it/s]

  0%|          | 0/2000 [00:00<?, ?it/s]

  0%|          | 0/2000 [00:00<?, ?it/s]

Sampling time =  0:00:15.309789
Transforming variables...
Transformation time =  0:00:00.289041
Computing Log Likelihood...

az.summary(idata_m2, var_names=["beta_ic", "beta_oc", "alpha"])

	mean	sd	hdi_3%	hdi_97%	mcse_mean	mcse_sd	ess_bulk	ess_tail	r_hat
beta_ic	0.001	0.000	-0.000	0.001	0.000	0.000	1215.0	1612.0	1.00
beta_oc	-0.003	0.001	-0.005	-0.001	0.000	0.000	1379.0	1743.0	1.00
alpha[ec]	1.039	0.497	0.076	1.936	0.016	0.012	908.0	1054.0	1.00
alpha[er]	1.077	0.474	0.216	1.988	0.016	0.012	839.0	991.0	1.00
alpha[gc]	2.376	0.309	1.789	2.953	0.011	0.008	814.0	836.0	1.01
alpha[gr]	0.733	0.373	-0.031	1.374	0.013	0.009	854.0	947.0	1.01

We can see now how this model performs much better in capturing aspects of the data generating process.

fig, axs = plt.subplots(1, 2, figsize=(20, 10))
ax = axs[0]
counts = wide_heating_df.groupby("depvar")["idcase"].count()
predicted_shares = idata_m2["posterior"]["p"].mean(dim=["chain", "draw", "obs"])
ci_lb = idata_m2["posterior"]["p"].quantile(0.025, dim=["chain", "draw", "obs"])
ci_ub = idata_m2["posterior"]["p"].quantile(0.975, dim=["chain", "draw", "obs"])

ax.scatter(ci_lb, ["ec", "er", "gc", "gr", "hp"], color="k", s=2)
ax.scatter(ci_ub, ["ec", "er", "gc", "gr", "hp"], color="k", s=2)
ax.scatter(
    counts / counts.sum(),
    ["ec", "er", "gc", "gr", "hp"],
    label="Observed Shares",
    color="red",
    s=100,
)
ax.hlines(
    ["ec", "er", "gc", "gr", "hp"], ci_lb, ci_ub, label="Predicted 95% Interval", color="black"
)
ax.legend()
ax.set_title("Observed V Predicted Shares")
az.plot_ppc(idata_m2, ax=axs[1])
axs[1].set_title("Posterior Predictive Checks")
ax.set_xlabel("Shares")
ax.set_ylabel("Heating System");

This model represents a substantial improvement.

Experimental Model: Adding Correlation Structure

We might think that there is a correlation among the alternative goods that we should capture too. We can capture those effects in so far as they exist by placing a multvariate normal prior on the intercepts, (or alternatively the beta parameters). In addition we add information about how the effect of income influences the utility accorded to each alternative.

coords = {
    "alts_intercepts": ["ec", "er", "gc", "gr"],
    "alts_probs": ["ec", "er", "gc", "gr", "hp"],
    "obs": range(N),
}
with pm.Model(coords=coords) as model_3:
    ## Add data to experiment with changes later.
    ic_ec = pm.MutableData("ic_ec", wide_heating_df["ic.ec"])
    oc_ec = pm.MutableData("oc_ec", wide_heating_df["oc.ec"])
    ic_er = pm.MutableData("ic_er", wide_heating_df["ic.er"])
    oc_er = pm.MutableData("oc_er", wide_heating_df["oc.er"])

    beta_ic = pm.Normal("beta_ic", 0, 1)
    beta_oc = pm.Normal("beta_oc", 0, 1)
    beta_income = pm.Normal("beta_income", 0, 1, dims="alts_intercepts")
    chol, corr, stds = pm.LKJCholeskyCov(
        "chol", n=4, eta=2.0, sd_dist=pm.Exponential.dist(1.0, shape=4)
    )
    alphas = pm.MvNormal("alpha", mu=0, chol=chol, dims="alts_intercepts")

    u0 = alphas[0] + beta_ic * ic_ec + beta_oc * oc_ec + beta_income[0] * wide_heating_df["income"]
    u1 = alphas[1] + beta_ic * ic_er + beta_oc * oc_er + beta_income[1] * wide_heating_df["income"]
    u2 = (
        alphas[2]
        + beta_ic * wide_heating_df["ic.gc"]
        + beta_oc * wide_heating_df["oc.gc"]
        + beta_income[2] * wide_heating_df["income"]
    )
    u3 = (
        alphas[3]
        + beta_ic * wide_heating_df["ic.gr"]
        + beta_oc * wide_heating_df["oc.gr"]
        + beta_income[3] * wide_heating_df["income"]
    )
    u4 = np.zeros(N)  # pivot
    s = pm.math.stack([u0, u1, u2, u3, u4]).T

    p_ = pm.Deterministic("p", pm.math.softmax(s, axis=1), dims=("obs", "alts_probs"))
    choice_obs = pm.Categorical("y_cat", p=p_, observed=observed, dims="obs")

    idata_m3 = pm.sample_prior_predictive()
    idata_m3.extend(
        pm.sample(nuts_sampler="numpyro", idata_kwargs={"log_likelihood": True}, random_seed=100)
    )
    idata_m3.extend(pm.sample_posterior_predictive(idata_m3))


pm.model_to_graphviz(model_3)

Sampling: [alpha, beta_ic, beta_income, beta_oc, chol, y_cat]
/Users/nathanielforde/mambaforge/envs/pymc_examples_new/lib/python3.9/site-packages/pymc/sampling/mcmc.py:243: UserWarning: Use of external NUTS sampler is still experimental
  warnings.warn("Use of external NUTS sampler is still experimental", UserWarning)

Compiling...
Compilation time =  0:00:04.533953
Sampling...

  0%|          | 0/2000 [00:00<?, ?it/s]

  0%|          | 0/2000 [00:00<?, ?it/s]

  0%|          | 0/2000 [00:00<?, ?it/s]

  0%|          | 0/2000 [00:00<?, ?it/s]

Sampling time =  0:00:30.042792
Transforming variables...
Transformation time =  0:00:00.594139
Computing Log Likelihood...

Plotting the model fit we see a similar story.The model predictive performance is not drastically improved and we have added some complexity to the model. This extra complexity ought to be penalised in model assessment metrics such as AIC and WAIC. But often the correlation amongst products are some of the features of interest, independent of issues of historic predictions.

fig, axs = plt.subplots(1, 2, figsize=(20, 10))
ax = axs[0]
counts = wide_heating_df.groupby("depvar")["idcase"].count()
predicted_shares = idata_m3["posterior"]["p"].mean(dim=["chain", "draw", "obs"])
ci_lb = idata_m3["posterior"]["p"].quantile(0.025, dim=["chain", "draw", "obs"])
ci_ub = idata_m3["posterior"]["p"].quantile(0.975, dim=["chain", "draw", "obs"])

ax.scatter(ci_lb, ["ec", "er", "gc", "gr", "hp"], color="k", s=2)
ax.scatter(ci_ub, ["ec", "er", "gc", "gr", "hp"], color="k", s=2)
ax.scatter(
    counts / counts.sum(),
    ["ec", "er", "gc", "gr", "hp"],
    label="Observed Shares",
    color="red",
    s=100,
)
ax.hlines(
    ["ec", "er", "gc", "gr", "hp"], ci_lb, ci_ub, label="Predicted 95% Interval", color="black"
)
ax.legend()
ax.set_title("Observed V Predicted Shares")
az.plot_ppc(idata_m3, ax=axs[1])
axs[1].set_title("Posterior Predictive Checks")
ax.set_xlabel("Shares")
ax.set_ylabel("Heating System");

That extra complexity can be informative, and the degree of relationship amongst the alternative products will inform the substitution patterns under policy changes. Also, note how under this model specification the parameter for beta_ic has a expected value of 0. Suggestive perhaps of a resignation towards the reality of installation costs that doesn't change the utility metric one way or other after a decision to purchase.

az.summary(
    idata_m3, var_names=["beta_income", "beta_ic", "beta_oc", "alpha", "chol_corr"], round_to=4
)

/Users/nathanielforde/mambaforge/envs/pymc_examples_new/lib/python3.9/site-packages/arviz/stats/diagnostics.py:592: RuntimeWarning: invalid value encountered in double_scalars
  (between_chain_variance / within_chain_variance + num_samples - 1) / (num_samples)

	mean	sd	hdi_3%	hdi_97%	mcse_mean	mcse_sd	ess_bulk	ess_tail	r_hat
beta_income[ec]	0.0971	0.1074	-0.1025	0.3046	0.0035	0.0025	936.3265	1900.0530	1.0033
beta_income[er]	0.0655	0.1047	-0.1187	0.2695	0.0036	0.0025	839.1058	1613.9147	1.0017
beta_income[gc]	0.0673	0.0867	-0.1058	0.2202	0.0032	0.0023	722.9224	1321.0255	1.0028
beta_income[gr]	-0.0318	0.0977	-0.2220	0.1441	0.0034	0.0024	807.8161	1624.6096	1.0020
beta_ic	0.0004	0.0007	-0.0009	0.0016	0.0000	0.0000	752.9909	914.0799	1.0019
beta_oc	-0.0035	0.0015	-0.0064	-0.0007	0.0000	0.0000	1436.0405	2066.3187	1.0015
alpha[ec]	1.0354	1.0479	-0.4211	3.0541	0.0470	0.0333	520.2449	1178.1694	1.0063
alpha[er]	1.2391	1.0751	-0.3175	3.2426	0.0507	0.0358	441.6820	991.4928	1.0064
alpha[gc]	2.3718	0.7613	1.1220	3.7710	0.0366	0.0259	414.8905	699.3486	1.0073
alpha[gr]	1.2014	0.8524	-0.0952	2.8006	0.0402	0.0284	442.3913	1198.3044	1.0053
chol_corr[0, 0]	1.0000	0.0000	1.0000	1.0000	0.0000	0.0000	4000.0000	4000.0000	NaN
chol_corr[0, 1]	0.1184	0.3671	-0.5402	0.7923	0.0074	0.0062	2518.0052	2043.9328	1.0015
chol_corr[0, 2]	0.1427	0.3705	-0.5480	0.7769	0.0093	0.0066	1673.7845	1975.7307	1.0020
chol_corr[0, 3]	0.1157	0.3753	-0.5676	0.7683	0.0079	0.0056	2319.4753	2119.7780	1.0012
chol_corr[1, 0]	0.1184	0.3671	-0.5402	0.7923	0.0074	0.0062	2518.0052	2043.9328	1.0015
chol_corr[1, 1]	1.0000	0.0000	1.0000	1.0000	0.0000	0.0000	4239.6296	4000.0000	0.9996
chol_corr[1, 2]	0.1675	0.3483	-0.4430	0.8095	0.0079	0.0056	1978.9399	1538.2851	1.0011
chol_corr[1, 3]	0.1526	0.3561	-0.4722	0.7963	0.0070	0.0050	2595.1991	3126.5524	1.0014
chol_corr[2, 0]	0.1427	0.3705	-0.5480	0.7769	0.0093	0.0066	1673.7845	1975.7307	1.0020
chol_corr[2, 1]	0.1675	0.3483	-0.4430	0.8095	0.0079	0.0056	1978.9399	1538.2851	1.0011
chol_corr[2, 2]	1.0000	0.0000	1.0000	1.0000	0.0000	0.0000	3929.0431	4000.0000	1.0007
chol_corr[2, 3]	0.1757	0.3411	-0.4384	0.7867	0.0071	0.0051	2260.6724	2564.2728	1.0017
chol_corr[3, 0]	0.1157	0.3753	-0.5676	0.7683	0.0079	0.0056	2319.4753	2119.7780	1.0012
chol_corr[3, 1]	0.1526	0.3561	-0.4722	0.7963	0.0070	0.0050	2595.1991	3126.5524	1.0014
chol_corr[3, 2]	0.1757	0.3411	-0.4384	0.7867	0.0071	0.0051	2260.6724	2564.2728	1.0017
chol_corr[3, 3]	1.0000	0.0000	1.0000	1.0000	0.0000	0.0000	3954.4789	3702.0363	1.0001

In this model we see that the marginal rate of substitution shows that an increase of one dollar for the operating costs is almost 17 times more impactful on the utility calculus than a similar increase in installation costs. Which makes sense in so far as we can expect the installation costs to be a one-off expense we're pretty resigned to.

post = az.extract(idata_m3)
substitution_rate = post["beta_oc"] / post["beta_ic"]
substitution_rate.mean().item()

17.581376035151784

Market Inteventions and Predicting Market Share

We can additionally use these kinds of models to predict market share under interventions where we change the price offering.

with model_3:
    # update values of predictors with new 20% price increase in operating costs for electrical options
    pm.set_data({"oc_ec": wide_heating_df["oc.ec"] * 1.2, "oc_er": wide_heating_df["oc.er"] * 1.2})
    # use the updated values and predict outcomes and probabilities:
    idata_new_policy = pm.sample_posterior_predictive(
        idata_m3,
        var_names=["p", "y_cat"],
        return_inferencedata=True,
        predictions=True,
        extend_inferencedata=False,
        random_seed=100,
    )

idata_new_policy

arviz.InferenceData

• predictions

  <xarray.Dataset>
  Dimensions:     (chain: 4, draw: 1000, obs: 900, alts_probs: 5)
  Coordinates:
    * chain       (chain) int64 0 1 2 3
    * draw        (draw) int64 0 1 2 3 4 5 6 7 ... 992 993 994 995 996 997 998 999
    * obs         (obs) int64 0 1 2 3 4 5 6 7 ... 892 893 894 895 896 897 898 899
    * alts_probs  (alts_probs) <U2 'ec' 'er' 'gc' 'gr' 'hp'
  Data variables:
      p           (chain, draw, obs, alts_probs) float64 0.04832 ... 0.05605
      y_cat       (chain, draw, obs) int64 2 4 4 2 1 2 2 2 2 ... 2 3 3 3 2 1 2 4 2
  Attributes:
      created_at:                 2023-07-13T17:30:54.332027
      arviz_version:              0.15.1
      inference_library:          pymc
      inference_library_version:  5.3.0

  xarray.Dataset

  • Dimensions:
    • chain: 4
    • draw: 1000
    • obs: 900
    • alts_probs: 5
  • Coordinates: (4)
    • chain
      (chain)
      int64
      0 1 2 3

      array([0, 1, 2, 3])

    • draw
      (draw)
      int64
      0 1 2 3 4 5 ... 995 996 997 998 999

      array([  0,   1,   2, ..., 997, 998, 999])

    • obs
      (obs)
      int64
      0 1 2 3 4 5 ... 895 896 897 898 899

      array([  0,   1,   2, ..., 897, 898, 899])

    • alts_probs
      (alts_probs)
      <U2
      'ec' 'er' 'gc' 'gr' 'hp'

      array(['ec', 'er', 'gc', 'gr', 'hp'], dtype='<U2')

  • Data variables: (2)
    • p
      (chain, draw, obs, alts_probs)
      float64
      0.04832 0.07521 ... 0.1829 0.05605

      array([[[[0.04832002, 0.0752078 , 0.68683941, 0.12152489, 0.06810788],
               [0.04476203, 0.07039096, 0.68099044, 0.13176806, 0.07208852],
               [0.05135906, 0.08271831, 0.64560802, 0.14763801, 0.07267661],
               ...,
               [0.05014134, 0.07771716, 0.66259606, 0.13810671, 0.07143873],
               [0.05050141, 0.09423651, 0.637604  , 0.14221295, 0.07544514],
               [0.03778261, 0.06934743, 0.64020447, 0.16589454, 0.08677095]],
      
              [[0.04233181, 0.06143711, 0.7111071 , 0.12979504, 0.05532895],
               [0.03947445, 0.05478907, 0.70784896, 0.13031194, 0.06757557],
               [0.05063401, 0.07792033, 0.62771886, 0.17004312, 0.07368368],
               ...,
               [0.04599708, 0.06829343, 0.65431144, 0.17033189, 0.06106616],
               [0.04250738, 0.09758275, 0.62190542, 0.17427809, 0.06372637],
               [0.02895578, 0.05159981, 0.63847768, 0.20485859, 0.07610814]],
      
              [[0.03431323, 0.05354092, 0.73089291, 0.12294999, 0.05830295],
               [0.03771595, 0.04837363, 0.71414228, 0.12773264, 0.0720355 ],
               [0.05356146, 0.07272789, 0.6164005 , 0.17813399, 0.07917616],
               ...,
      ...
               ...,
               [0.05726533, 0.06342393, 0.66066572, 0.1516834 , 0.06696161],
               [0.05086387, 0.08283043, 0.64817392, 0.15153051, 0.06660127],
               [0.03979472, 0.07152408, 0.65481527, 0.17632445, 0.05754148]],
      
              [[0.05094682, 0.07403621, 0.67371889, 0.13510685, 0.06619123],
               [0.05190867, 0.0751801 , 0.66746741, 0.14668949, 0.05875433],
               [0.05890911, 0.08693934, 0.63771287, 0.161988  , 0.05445068],
               ...,
               [0.05589228, 0.08072595, 0.65194827, 0.15355613, 0.05787737],
               [0.0557648 , 0.09253297, 0.63544208, 0.15638201, 0.05987814],
               [0.05083981, 0.08123078, 0.63433452, 0.1824528 , 0.0511421 ]],
      
              [[0.05265113, 0.05813257, 0.68305352, 0.15017725, 0.05598552],
               [0.04728697, 0.05506159, 0.68930742, 0.15935793, 0.0489861 ],
               [0.05892811, 0.06923376, 0.65607418, 0.1716434 , 0.04412055],
               ...,
               [0.05729577, 0.06241275, 0.67345633, 0.15633063, 0.05050452],
               [0.06081723, 0.0825264 , 0.63398358, 0.16625131, 0.05642149],
               [0.03967079, 0.05927182, 0.66208718, 0.18291914, 0.05605107]]]])

    • y_cat
      (chain, draw, obs)
      int64
      2 4 4 2 1 2 2 2 ... 3 3 3 2 1 2 4 2

      array([[[2, 4, 4, ..., 3, 2, 2],
              [2, 0, 0, ..., 2, 2, 2],
              [2, 2, 2, ..., 2, 2, 2],
              ...,
              [3, 2, 1, ..., 2, 2, 2],
              [2, 2, 2, ..., 2, 2, 2],
              [0, 2, 2, ..., 2, 2, 2]],
      
             [[0, 2, 2, ..., 2, 0, 2],
              [2, 0, 2, ..., 2, 0, 1],
              [4, 2, 2, ..., 2, 3, 4],
              ...,
              [2, 2, 2, ..., 2, 2, 2],
              [2, 3, 2, ..., 3, 0, 2],
              [2, 2, 2, ..., 2, 2, 2]],
      
             [[2, 2, 1, ..., 2, 2, 3],
              [2, 2, 4, ..., 1, 0, 1],
              [2, 0, 4, ..., 2, 4, 3],
              ...,
              [4, 4, 2, ..., 2, 1, 2],
              [2, 4, 2, ..., 2, 2, 2],
              [3, 3, 3, ..., 2, 4, 2]],
      
             [[2, 2, 3, ..., 1, 2, 2],
              [2, 2, 2, ..., 2, 2, 1],
              [2, 2, 2, ..., 2, 2, 2],
              ...,
              [2, 0, 2, ..., 2, 1, 2],
              [2, 2, 2, ..., 3, 2, 3],
              [2, 3, 1, ..., 2, 4, 2]]])

  • Indexes: (4)
    • chain
      PandasIndex

      PandasIndex(Int64Index([0, 1, 2, 3], dtype='int64', name='chain'))

    • draw
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  990, 991, 992, 993, 994, 995, 996, 997, 998, 999],
                 dtype='int64', name='draw', length=1000))

    • obs
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  890, 891, 892, 893, 894, 895, 896, 897, 898, 899],
                 dtype='int64', name='obs', length=900))

    • alts_probs
      PandasIndex

      PandasIndex(Index(['ec', 'er', 'gc', 'gr', 'hp'], dtype='object', name='alts_probs'))

  • Attributes: (4)

    created_at :

    2023-07-13T17:30:54.332027

    arviz_version :

    0.15.1

    inference_library :

    pymc

    inference_library_version :

    5.3.0

  
  
• predictions_constant_data

  <xarray.Dataset>
  Dimensions:      (ic_ec_dim_0: 900, oc_ec_dim_0: 900, ic_er_dim_0: 900,
                    oc_er_dim_0: 900)
  Coordinates:
    * ic_ec_dim_0  (ic_ec_dim_0) int64 0 1 2 3 4 5 6 ... 894 895 896 897 898 899
    * oc_ec_dim_0  (oc_ec_dim_0) int64 0 1 2 3 4 5 6 ... 894 895 896 897 898 899
    * ic_er_dim_0  (ic_er_dim_0) int64 0 1 2 3 4 5 6 ... 894 895 896 897 898 899
    * oc_er_dim_0  (oc_er_dim_0) int64 0 1 2 3 4 5 6 ... 894 895 896 897 898 899
  Data variables:
      ic_ec        (ic_ec_dim_0) float64 859.9 796.8 719.9 ... 842.8 799.8 967.9
      oc_ec        (oc_ec_dim_0) float64 664.0 624.3 526.9 ... 574.6 594.2 622.4
      ic_er        (ic_er_dim_0) float64 995.8 894.7 900.1 ... 1.123e+03 1.092e+03
      oc_er        (oc_er_dim_0) float64 606.7 583.8 485.7 ... 538.3 481.9 550.2
  Attributes:
      created_at:                 2023-07-13T17:30:54.333501
      arviz_version:              0.15.1
      inference_library:          pymc
      inference_library_version:  5.3.0

  xarray.Dataset

  • Dimensions:
    • ic_ec_dim_0: 900
    • oc_ec_dim_0: 900
    • ic_er_dim_0: 900
    • oc_er_dim_0: 900
  • Coordinates: (4)
    • ic_ec_dim_0
      (ic_ec_dim_0)
      int64
      0 1 2 3 4 5 ... 895 896 897 898 899

      array([  0,   1,   2, ..., 897, 898, 899])

    • oc_ec_dim_0
      (oc_ec_dim_0)
      int64
      0 1 2 3 4 5 ... 895 896 897 898 899

      array([  0,   1,   2, ..., 897, 898, 899])

    • ic_er_dim_0
      (ic_er_dim_0)
      int64
      0 1 2 3 4 5 ... 895 896 897 898 899

      array([  0,   1,   2, ..., 897, 898, 899])

    • oc_er_dim_0
      (oc_er_dim_0)
      int64
      0 1 2 3 4 5 ... 895 896 897 898 899

      array([  0,   1,   2, ..., 897, 898, 899])

  • Data variables: (4)
    • ic_ec
      (ic_ec_dim_0)
      float64
      859.9 796.8 719.9 ... 799.8 967.9

      array([ 859.9 ,  796.82,  719.86,  761.25,  858.86,  693.74,  633.63,
              813.33,  875.7 ,  775.86,  792.29,  944.35,  797.34,  687.18,
              594.  ,  801.2 ,  821.79,  696.89,  572.3 ,  925.62,  788.07,
              934.75,  619.65,  805.29,  943.05,  882.55,  869.65,  991.82,
              724.01,  805.08,  838.01,  671.78, 1030.3 ,  785.95,  689.1 ,
              918.92,  618.3 ,  891.58,  899.  ,  776.67,  788.79,  891.47,
              828.73,  874.67,  796.43,  938.58,  588.55,  825.11,  828.27,
              733.4 ,  754.26,  836.85,  963.76,  751.67, 1021.7 ,  843.18,
              625.26,  970.49,  727.13,  762.51,  747.32,  608.57,  937.96,
              912.46,  805.94,  790.89,  743.31,  702.5 ,  901.4 ,  971.55,
              612.24,  780.07,  776.18,  804.51,  683.38,  994.11,  696.87,
              680.51,  830.42,  829.45,  721.85,  602.71,  845.69,  798.54,
              757.17,  800.49,  890.49,  850.44,  834.43,  573.53,  689.92,
              704.06,  692.2 ,  993.06,  948.83,  826.91,  700.09,  763.54,
              977.49,  772.38, 1034.7 ,  857.21,  773.6 ,  941.68,  780.37,
              782.98,  751.48,  832.19,  830.27,  889.23,  830.64,  746.92,
              804.48,  682.43,  558.6 ,  781.41,  959.49,  816.38,  860.01,
              702.16, 1115.3 ,  801.58,  867.54,  837.86,  792.1 , 1011.1 ,
              959.82,  829.68,  825.94,  699.84,  875.86,  835.  ,  889.76,
              746.42,  769.34,  686.17,  638.89,  968.98,  746.87,  951.02,
      ...
             1088.9 ,  953.25,  975.47,  739.3 ,  683.59,  762.3 ,  829.64,
              678.08, 1132.1 ,  927.19,  850.24,  920.17,  871.7 ,  923.98,
              724.22,  634.18,  778.44,  829.65,  653.52, 1084.7 ,  866.22,
             1104.4 ,  747.36,  802.2 ,  655.31,  745.85,  775.23,  777.17,
              974.98,  861.16,  542.66,  924.5 ,  583.76,  944.95,  708.19,
              711.34,  904.8 ,  832.63,  736.37,  755.84,  932.63,  877.49,
              835.05,  894.63,  738.95,  729.29,  728.25,  727.56,  856.33,
             1170.2 ,  677.33,  769.68,  917.17,  729.99,  906.88,  787.11,
              758.53,  891.55, 1056.2 ,  765.5 ,  935.82,  654.82, 1060.4 ,
              904.44,  754.59,  877.16,  705.33,  907.46, 1035.7 ,  778.12,
              952.24,  987.31,  710.32,  813.24,  915.55,  832.4 ,  690.63,
              924.76,  782.23, 1052.2 ,  915.47,  591.07,  580.45,  654.78,
              951.81,  849.29,  752.82,  657.91,  645.62,  818.47,  925.21,
              711.15,  977.86,  884.21,  814.93,  790.7 ,  832.93,  869.97,
              772.97,  861.66,  971.8 ,  774.22,  717.59, 1068.1 ,  902.09,
              779.6 ,  956.44,  890.78,  783.42,  863.74,  984.2 ,  689.48,
              896.96,  838.66,  788.  ,  923.13,  940.1 ,  993.93,  711.93,
              722.87,  805.08,  960.95,  762.37,  916.23,  758.47,  557.39,
              982.54,  898.36,  834.27,  822.31,  984.25,  813.69,  751.59,
             1047.6 ,  842.79,  799.76,  967.88])

    • oc_ec
      (oc_ec_dim_0)
      float64
      664.0 624.3 526.9 ... 594.2 622.4

      array([664.008, 624.288, 526.872, 579.6  , 485.292, 477.864, 574.032,
             602.88 , 663.576, 638.664, 513.72 , 681.996, 488.052, 439.104,
             509.652, 531.792, 519.096, 435.588, 409.392, 674.028, 529.188,
             596.124, 500.784, 718.98 , 692.64 , 633.456, 584.388, 701.796,
             532.584, 648.384, 623.124, 506.784, 720.696, 642.576, 503.52 ,
             566.688, 537.24 , 591.828, 499.128, 589.32 , 632.616, 708.684,
             518.904, 521.532, 546.096, 584.244, 463.596, 565.524, 656.28 ,
             564.276, 501.384, 623.688, 671.916, 659.772, 647.748, 527.64 ,
             499.392, 712.848, 474.24 , 524.52 , 550.26 , 473.448, 666.216,
             585.408, 628.068, 578.172, 490.488, 473.796, 697.344, 586.788,
             458.232, 579.78 , 484.992, 523.872, 439.368, 648.528, 468.072,
             574.368, 648.804, 580.5  , 541.224, 518.736, 565.44 , 588.36 ,
             557.472, 589.644, 530.724, 496.536, 504.312, 493.68 , 599.184,
             589.488, 595.86 , 620.028, 728.496, 672.396, 522.744, 484.932,
             667.8  , 632.736, 662.784, 472.596, 579.276, 604.608, 546.744,
             583.536, 512.172, 470.628, 615.288, 682.608, 669.216, 528.156,
             562.332, 424.152, 505.056, 587.712, 664.548, 613.416, 636.456,
             654.984, 758.172, 578.304, 575.04 , 576.504, 560.592, 780.384,
             490.992, 586.86 , 582.9  , 567.3  , 607.98 , 499.668, 536.76 ,
             674.22 , 518.712, 598.872, 511.092, 561.876, 547.236, 557.58 ,
      ...
             729.12 , 691.02 , 647.256, 591.   , 585.54 , 486.444, 630.516,
             576.912, 688.512, 652.884, 651.06 , 603.876, 644.4  , 616.896,
             511.668, 427.152, 642.18 , 594.648, 456.348, 706.14 , 600.468,
             796.368, 486.396, 613.932, 439.536, 506.604, 510.324, 414.024,
             658.32 , 683.436, 380.7  , 517.08 , 423.552, 569.496, 441.888,
             603.576, 580.38 , 668.256, 531.972, 580.44 , 475.8  , 644.988,
             624.624, 664.44 , 399.588, 509.172, 434.292, 409.38 , 522.444,
             814.404, 496.116, 589.176, 683.58 , 500.172, 597.876, 531.948,
             449.4  , 660.204, 735.396, 658.956, 593.208, 386.94 , 661.584,
             610.68 , 394.104, 590.472, 597.36 , 566.676, 609.06 , 569.16 ,
             603.168, 614.148, 539.172, 477.228, 557.292, 513.18 , 479.64 ,
             709.668, 596.532, 743.748, 572.244, 466.632, 406.632, 488.16 ,
             632.676, 639.624, 494.016, 505.128, 514.92 , 539.1  , 647.304,
             471.324, 634.2  , 614.928, 585.468, 626.712, 562.776, 677.52 ,
             604.272, 658.152, 643.668, 474.516, 558.552, 773.664, 704.196,
             508.932, 720.432, 539.436, 588.96 , 642.936, 613.992, 522.372,
             603.24 , 596.772, 569.58 , 679.92 , 683.508, 602.784, 538.188,
             630.816, 524.808, 648.192, 568.248, 531.756, 609.78 , 446.16 ,
             598.536, 603.384, 481.284, 590.22 , 653.82 , 608.268, 569.376,
             846.432, 574.632, 594.24 , 622.416])

    • ic_er
      (ic_er_dim_0)
      float64
      995.8 894.7 ... 1.123e+03 1.092e+03

      array([ 995.76,  894.69,  900.11,  831.04,  985.64,  862.56,  952.3 ,
             1012.3 , 1024.8 ,  874.49,  774.69, 1325.  ,  880.8 ,  718.72,
              736.77, 1092.4 ,  969.64,  853.82,  664.22, 1140.  ,  799.52,
             1009.5 ,  880.97,  806.59, 1130.5 , 1187.9 ,  942.49,  913.32,
              993.25,  941.41, 1144.4 ,  976.01, 1153.2 ,  963.54,  876.97,
              897.8 ,  819.38, 1070.1 , 1060.2 , 1054.  ,  833.05, 1148.3 ,
              938.83,  966.34, 1054.4 , 1098.1 ,  810.51, 1143.9 ,  929.43,
             1012.7 ,  994.12, 1026.5 , 1278.3 ,  906.09, 1162.7 ,  884.68,
              736.03, 1330.1 ,  878.61,  996.61, 1130.9 ,  810.97, 1100.5 ,
              950.35, 1304.  , 1115.9 ,  856.94,  752.47, 1197.4 , 1043.9 ,
              795.83,  946.68,  824.73,  831.37,  865.94, 1081.8 ,  916.09,
             1191.8 , 1059.3 , 1159.6 ,  921.97,  790.37,  866.67,  948.05,
              903.52, 1051.1 ,  923.99,  990.48,  845.59,  670.05,  964.66,
             1131.4 ,  848.99, 1075.7 , 1031.8 ,  993.83,  856.06,  980.45,
             1290.7 ,  862.07, 1185.4 , 1072.5 , 1023.6 ,  934.18,  902.04,
             1039.1 ,  909.56,  961.74,  963.76, 1107.1 ,  946.4 ,  917.13,
              822.85,  712.86,  714.01,  966.06, 1146.7 , 1120.  , 1048.7 ,
              984.69, 1225.5 , 1125.4 , 1004.8 , 1027.6 , 1015.1 , 1142.5 ,
              937.12,  879.88,  949.07,  875.49,  942.76,  957.79, 1028.2 ,
             1143.9 ,  979.5 ,  931.27,  722.43,  942.23,  851.91, 1082.1 ,
      ...
             1210.  , 1168.1 , 1210.7 ,  989.9 ,  935.29,  824.84, 1015.9 ,
             1121.  , 1174.5 , 1089.5 , 1116.4 , 1050.6 ,  910.56, 1069.6 ,
              993.27,  777.28, 1150.1 , 1038.7 ,  911.5 , 1275.4 ,  906.27,
             1416.7 ,  845.54, 1003.1 ,  891.1 ,  919.38,  854.6 ,  825.66,
             1122.1 , 1022.7 ,  835.91, 1084.8 ,  658.38, 1147.2 ,  757.77,
              794.42,  933.16,  965.96,  893.67,  912.42, 1123.3 , 1140.7 ,
              894.26, 1173.7 ,  762.06,  859.23,  803.77,  825.  , 1131.5 ,
             1417.8 ,  912.44,  766.79, 1034.7 ,  653.97, 1183.1 ,  865.21,
              704.74, 1012.5 , 1263.4 , 1057.5 , 1023.5 ,  622.99, 1304.7 ,
             1078.2 ,  972.28, 1050.2 ,  907.67,  907.25, 1017.8 ,  985.2 ,
              891.76, 1157.5 ,  910.15,  809.95, 1105.8 ,  993.83,  890.9 ,
             1175.6 ,  979.25, 1293.8 , 1026.3 ,  733.76,  777.04,  849.56,
             1138.7 , 1255.9 ,  845.15,  843.31,  931.4 , 1029.8 , 1165.4 ,
              871.97,  963.42, 1156.5 , 1150.3 , 1058.  , 1080.6 , 1080.4 ,
             1042.4 ,  973.06,  943.57,  865.61,  863.93, 1328.  , 1057.7 ,
              988.55, 1132.  ,  989.09,  685.23,  980.36, 1121.1 ,  840.17,
              966.41, 1175.4 , 1117.6 , 1265.8 , 1148.6 , 1067.5 ,  972.16,
             1169.7 ,  996.96,  965.23,  907.4 , 1067.7 , 1057.3 ,  650.59,
             1042.7 ,  988.99,  987.05, 1045.3 , 1246.5 , 1064.1 ,  869.78,
             1292.6 , 1041.3 , 1123.2 , 1091.7 ])

    • oc_er
      (oc_er_dim_0)
      float64
      606.7 583.8 485.7 ... 481.9 550.2

      array([606.72 , 583.788, 485.688, 510.264, 467.424, 445.248, 535.164,
             557.724, 542.112, 566.868, 456.936, 570.864, 405.636, 474.168,
             387.54 , 456.228, 535.656, 538.116, 348.888, 633.864, 343.548,
             497.748, 434.688, 467.52 , 612.792, 546.96 , 512.532, 567.72 ,
             528.252, 523.596, 596.028, 494.952, 660.264, 567.252, 438.192,
             505.908, 410.832, 523.5  , 568.38 , 519.096, 480.552, 696.876,
             479.748, 525.684, 491.664, 559.428, 355.692, 558.72 , 582.048,
             613.944, 517.752, 446.76 , 679.86 , 421.752, 584.82 , 451.968,
             439.788, 712.812, 479.844, 578.652, 409.14 , 489.132, 512.1  ,
             558.   , 555.012, 505.428, 471.924, 331.92 , 613.68 , 533.052,
             408.072, 489.732, 492.936, 560.292, 413.088, 494.928, 433.2  ,
             438.852, 420.168, 503.172, 471.684, 353.136, 515.94 , 569.664,
             509.472, 536.46 , 464.484, 477.768, 472.416, 460.572, 474.168,
             553.368, 497.496, 548.724, 586.32 , 574.2  , 526.404, 476.676,
             595.32 , 479.184, 634.272, 454.764, 546.144, 531.924, 479.784,
             536.76 , 517.584, 475.296, 565.344, 587.472, 521.268, 439.812,
             423.348, 284.004, 468.744, 376.212, 513.96 , 574.764, 574.32 ,
             484.908, 702.132, 530.832, 535.032, 507.372, 479.064, 645.84 ,
             483.972, 395.472, 499.956, 450.276, 516.864, 520.956, 516.468,
             556.884, 501.432, 513.684, 377.7  , 557.376, 499.932, 614.1  ,
      ...
             636.432, 585.42 , 673.332, 477.252, 486.312, 374.568, 479.928,
             534.396, 553.068, 605.112, 541.692, 519.756, 545.328, 544.38 ,
             526.176, 438.552, 593.088, 598.932, 394.584, 797.316, 503.556,
             685.86 , 449.736, 561.684, 422.76 , 477.42 , 528.276, 447.54 ,
             683.868, 577.956, 402.012, 464.784, 424.884, 516.432, 389.16 ,
             461.448, 444.012, 521.904, 427.008, 509.496, 449.28 , 589.416,
             458.88 , 532.44 , 506.16 , 449.652, 400.548, 357.744, 576.   ,
             730.716, 529.872, 478.98 , 603.684, 340.56 , 562.344, 542.436,
             422.58 , 530.268, 651.444, 427.308, 650.796, 440.52 , 707.772,
             561.96 , 406.44 , 497.232, 501.156, 480.876, 551.424, 563.46 ,
             565.5  , 619.176, 517.68 , 548.172, 547.332, 502.956, 437.364,
             598.656, 513.252, 638.868, 573.624, 369.432, 389.784, 398.76 ,
             568.212, 606.876, 491.568, 377.436, 509.028, 475.704, 656.064,
             400.752, 458.664, 673.512, 527.628, 446.244, 505.344, 549.672,
             655.86 , 561.024, 587.856, 487.74 , 412.308, 645.456, 624.312,
             436.968, 642.864, 572.004, 484.524, 481.344, 528.672, 440.82 ,
             495.936, 529.992, 496.212, 666.48 , 442.62 , 605.712, 531.216,
             414.684, 477.432, 506.34 , 556.272, 522.432, 556.464, 422.532,
             512.808, 524.892, 403.812, 546.756, 587.34 , 491.112, 504.78 ,
             661.932, 538.332, 481.872, 550.236])

  • Indexes: (4)
    • ic_ec_dim_0
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  890, 891, 892, 893, 894, 895, 896, 897, 898, 899],
                 dtype='int64', name='ic_ec_dim_0', length=900))

    • oc_ec_dim_0
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  890, 891, 892, 893, 894, 895, 896, 897, 898, 899],
                 dtype='int64', name='oc_ec_dim_0', length=900))

    • ic_er_dim_0
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  890, 891, 892, 893, 894, 895, 896, 897, 898, 899],
                 dtype='int64', name='ic_er_dim_0', length=900))

    • oc_er_dim_0
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  890, 891, 892, 893, 894, 895, 896, 897, 898, 899],
                 dtype='int64', name='oc_er_dim_0', length=900))

  • Attributes: (4)

    created_at :

    2023-07-13T17:30:54.333501

    arviz_version :

    0.15.1

    inference_library :

    pymc

    inference_library_version :

    5.3.0

  
  

idata_new_policy["predictions"]["p"].mean(dim=["chain", "draw", "obs"])

<xarray.DataArray 'p' (alts_probs: 5)>
array([0.05383866, 0.07239016, 0.66253495, 0.1482966 , 0.06293963])
Coordinates:
  * alts_probs  (alts_probs) <U2 'ec' 'er' 'gc' 'gr' 'hp'

xarray.DataArray

'p'

• alts_probs: 5

• 0.05384 0.07239 0.6625 0.1483 0.06294

  array([0.05383866, 0.07239016, 0.66253495, 0.1482966 , 0.06293963])

• Coordinates: (1)
  • alts_probs
    (alts_probs)
    <U2
    'ec' 'er' 'gc' 'gr' 'hp'

    array(['ec', 'er', 'gc', 'gr', 'hp'], dtype='<U2')

• Indexes: (1)
  • alts_probs
    PandasIndex

    PandasIndex(Index(['ec', 'er', 'gc', 'gr', 'hp'], dtype='object', name='alts_probs'))

• Attributes: (0)

fig, ax = plt.subplots(1, figsize=(20, 10))
counts = wide_heating_df.groupby("depvar")["idcase"].count()
new_predictions = idata_new_policy["predictions"]["p"].mean(dim=["chain", "draw", "obs"]).values
ci_lb = idata_m3["posterior"]["p"].quantile(0.025, dim=["chain", "draw", "obs"])
ci_ub = idata_m3["posterior"]["p"].quantile(0.975, dim=["chain", "draw", "obs"])
ax.scatter(ci_lb, ["ec", "er", "gc", "gr", "hp"], color="k", s=2)
ax.scatter(ci_ub, ["ec", "er", "gc", "gr", "hp"], color="k", s=2)
ax.scatter(
    new_predictions,
    ["ec", "er", "gc", "gr", "hp"],
    color="green",
    label="New Policy Predicted Share",
    s=100,
)
ax.scatter(
    counts / counts.sum(),
    ["ec", "er", "gc", "gr", "hp"],
    label="Observed Shares",
    color="red",
    s=100,
)
ax.hlines(
    ["ec", "er", "gc", "gr", "hp"],
    ci_lb,
    ci_ub,
    label="Predicted 95% Credible Interval Old Policy",
    color="black",
)
ax.set_title("Predicted Market Shares under Old and New Pricing Policy", fontsize=20)
ax.set_xlabel("Market Share")
ax.legend();

Here we can, as expected, see that a rise in the operating costs of the electrical options has a negative impact on their predicted market share.

Compare Models

We'll now evaluate all three model fits on their predictive performance. Predictive performance on the original data is a good benchmark that the model has appropriately captured the data generating process. But it is not (as we've seen) the only feature of interest in these models. These models are sensetive to our theoretical beliefs about the agents making the decisions, the view of the decision process and the elements of the choice scenario.

compare = az.compare({"m1": idata_m1, "m2": idata_m2, "m3": idata_m3})
compare

	rank	elpd_loo	p_loo	elpd_diff	weight	se	dse	warning	scale
m2	0	-1023.600927	4.964862	0.000000	1.000000e+00	27.802379	0.000000	False	log
m3	1	-1025.830780	9.954792	2.229854	2.220446e-16	28.086804	2.070976	False	log
m1	2	-1309.610895	1.196878	286.009968	0.000000e+00	12.933024	22.677606	False	log

az.plot_compare(compare)

<Axes: title={'center': 'Model comparison\nhigher is better'}, xlabel='elpd_loo (log)', ylabel='ranked models'>

Choosing Crackers over Repeated Choices: Mixed Logit Model

Moving to another example, we see a choice scenario where the same individual has been repeatedly polled on their choice of crackers among alternatives. This affords us the opportunity to evaluate the preferences of individuals by adding in coefficients for individuals for each product.

try:
    c_df = pd.read_csv("../data/cracker_choice_short.csv")
except:
    c_df = pd.read_csv(pm.get_data("cracker_choice_short.csv"))
columns = [c for c in c_df.columns if c != "Unnamed: 0"]
c_df[columns]

	personId	disp.sunshine	disp.keebler	disp.nabisco	disp.private	feat.sunshine	feat.keebler	feat.nabisco	feat.private	price.sunshine	price.keebler	price.nabisco	price.private	choice	lastChoice	personChoiceId	choiceId
0	1	0	0	0	0	0	0	0	0	0.99	1.09	0.99	0.71	nabisco	nabisco	1	1
1	1	1	0	0	0	0	0	0	0	0.49	1.09	1.09	0.78	sunshine	nabisco	2	2
2	1	0	0	0	0	0	0	0	0	1.03	1.09	0.89	0.78	nabisco	sunshine	3	3
3	1	0	0	0	0	0	0	0	0	1.09	1.09	1.19	0.64	nabisco	nabisco	4	4
4	1	0	0	0	0	0	0	0	0	0.89	1.09	1.19	0.84	nabisco	nabisco	5	5
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
3151	136	0	0	0	0	0	0	0	0	1.09	1.19	0.99	0.55	private	private	9	3152
3152	136	0	0	0	1	0	0	0	0	0.78	1.35	1.04	0.65	private	private	10	3153
3153	136	0	0	0	0	0	0	0	0	1.09	1.17	1.29	0.59	private	private	11	3154
3154	136	0	0	0	0	0	0	0	0	1.09	1.22	1.29	0.59	private	private	12	3155
3155	136	0	0	0	0	0	0	0	0	1.29	1.04	1.23	0.59	private	private	13	3156

3156 rows × 17 columns

c_df.groupby("personId")[["choiceId"]].count().T

personId	1	2	3	4	5	6	7	8	9	10	...	127	128	129	130	131	132	133	134	135	136
choiceId	15	15	13	28	13	27	16	25	18	40	...	17	25	31	31	29	21	26	13	14	13

1 rows × 136 columns

The presence of repeated choice over time complicates the issue. We now have to contend with issues of personal taste and the evolving or dynamic effects of pricing in a competitive environment. Plotting the simple linear and polynomial fits for each person's successive exposure to the brand price, seems to suggest that (a) pricing differentiates the product offering and (b) pricing evolves over time.

fig, axs = plt.subplots(1, 2, figsize=(20, 10))
axs = axs.flatten()
map_color = {"nabisco": "red", "keebler": "blue", "sunshine": "purple", "private": "orange"}


for i in c_df["personId"].unique():
    temp = c_df[c_df["personId"] == i].copy(deep=True)
    temp["color"] = temp["choice"].map(map_color)
    predict = np.poly1d(np.polyfit(temp["personChoiceId"], temp["price.sunshine"], deg=1))
    axs[0].plot(predict(range(25)), color="red", label="Sunshine", alpha=0.4)
    predict = np.poly1d(np.polyfit(temp["personChoiceId"], temp["price.keebler"], deg=1))
    axs[0].plot(predict(range(25)), color="blue", label="Keebler", alpha=0.4)
    predict = np.poly1d(np.polyfit(temp["personChoiceId"], temp["price.nabisco"], deg=1))
    axs[0].plot(predict(range(25)), color="grey", label="Nabisco", alpha=0.4)

    predict = np.poly1d(np.polyfit(temp["personChoiceId"], temp["price.sunshine"], deg=2))
    axs[1].plot(predict(range(25)), color="red", label="Sunshine", alpha=0.4)
    predict = np.poly1d(np.polyfit(temp["personChoiceId"], temp["price.keebler"], deg=2))
    axs[1].plot(predict(range(25)), color="blue", label="Keebler", alpha=0.4)
    predict = np.poly1d(np.polyfit(temp["personChoiceId"], temp["price.nabisco"], deg=2))
    axs[1].plot(predict(range(25)), color="grey", label="Nabisco", alpha=0.4)

axs[0].set_title("Linear Regression Fit \n Customer Price Exposure over Time", fontsize=20)
axs[1].set_title("Polynomial^(2) Regression Fit \n Customer Price Exposure over Time", fontsize=20)
axs[0].set_xlabel("Nth Decision/Time point")
axs[1].set_xlabel("Nth Decision/Time point")
axs[0].set_ylabel("Product Price Offered")
axs[1].set_ylim(0, 2)
axs[0].set_ylim(0, 2)

colors = ["red", "blue", "grey"]
lines = [Line2D([0], [0], color=c, linewidth=3, linestyle="-") for c in colors]
labels = ["Sunshine", "Keebler", "Nabisco"]
axs[0].legend(lines, labels)
axs[1].legend(lines, labels);

We'll model now how individual taste enters into discrete choice problems, but ignore the complexities of the time-dimension or the endogenity of price in the system. There are adaptions of the basic discrete choice model that are designed to address each of these complications. We'll leave the temporal dynamics as a suggested exercise for the reader.

N = c_df.shape[0]
observed = pd.Categorical(c_df["choice"]).codes
person_indx, uniques = pd.factorize(c_df["personId"])

coords = {
    "alts_intercepts": ["sunshine", "keebler", "nabisco"],
    "alts_probs": ["sunshine", "keebler", "nabisco", "private"],
    "individuals": uniques,
    "obs": range(N),
}
with pm.Model(coords=coords) as model_4:
    beta_feat = pm.TruncatedNormal("beta_feat", 0, 1, upper=10, lower=0)
    beta_disp = pm.TruncatedNormal("beta_disp", 0, 1, upper=10, lower=0)
    ## Stronger Prior on Price to ensure an increase in price negatively impacts utility
    beta_price = pm.TruncatedNormal("beta_price", 0, 1, upper=0, lower=-10)
    alphas = pm.Normal("alpha", 0, 1, dims="alts_intercepts")
    beta_individual = pm.Normal("beta_individual", 0, 0.05, dims=("individuals", "alts_intercepts"))

    u0 = (
        (alphas[0] + beta_individual[person_indx, 0])
        + beta_disp * c_df["disp.sunshine"]
        + beta_feat * c_df["feat.sunshine"]
        + beta_price * c_df["price.sunshine"]
    )
    u1 = (
        (alphas[1] + beta_individual[person_indx, 1])
        + beta_disp * c_df["disp.keebler"]
        + beta_feat * c_df["feat.keebler"]
        + beta_price * c_df["price.keebler"]
    )
    u2 = (
        (alphas[2] + beta_individual[person_indx, 2])
        + beta_disp * c_df["disp.nabisco"]
        + beta_feat * c_df["feat.nabisco"]
        + beta_price * c_df["price.nabisco"]
    )
    u3 = np.zeros(N)  # Outside Good
    s = pm.math.stack([u0, u1, u2, u3]).T
    # Reconstruct the total data

    ## Apply Softmax Transform
    p_ = pm.Deterministic("p", pm.math.softmax(s, axis=1), dims=("obs", "alts_probs"))

    ## Likelihood
    choice_obs = pm.Categorical("y_cat", p=p_, observed=observed, dims="obs")

    idata_m4 = pm.sample_prior_predictive()
    idata_m4.extend(
        pm.sample(nuts_sampler="numpyro", idata_kwargs={"log_likelihood": True}, random_seed=103)
    )
    idata_m4.extend(pm.sample_posterior_predictive(idata_m4))


pm.model_to_graphviz(model_4)

Sampling: [alpha, beta_disp, beta_feat, beta_individual, beta_price, y_cat]
/Users/nathanielforde/mambaforge/envs/pymc_examples_new/lib/python3.9/site-packages/pymc/sampling/mcmc.py:243: UserWarning: Use of external NUTS sampler is still experimental
  warnings.warn("Use of external NUTS sampler is still experimental", UserWarning)

Compiling...
Compilation time =  0:00:02.628050
Sampling...

  0%|          | 0/2000 [00:00<?, ?it/s]

  0%|          | 0/2000 [00:00<?, ?it/s]

  0%|          | 0/2000 [00:00<?, ?it/s]

  0%|          | 0/2000 [00:00<?, ?it/s]

Sampling time =  0:00:11.040460
Transforming variables...
Transformation time =  0:00:01.419801
Computing Log Likelihood...

az.summary(idata_m4, var_names=["beta_disp", "beta_feat", "beta_price", "alpha", "beta_individual"])

	mean	sd	hdi_3%	hdi_97%	mcse_mean	mcse_sd	ess_bulk	ess_tail	r_hat
beta_disp	0.023	0.021	0.000	0.061	0.000	0.000	5386.0	2248.0	1.0
beta_feat	0.019	0.018	0.000	0.052	0.000	0.000	6417.0	2283.0	1.0
beta_price	-0.021	0.020	-0.059	-0.000	0.000	0.000	4281.0	2277.0	1.0
alpha[sunshine]	-0.054	0.096	-0.243	0.119	0.002	0.001	3632.0	3220.0	1.0
alpha[keebler]	2.023	0.073	1.886	2.159	0.001	0.001	3302.0	2894.0	1.0
...	...	...	...	...	...	...	...	...	...
beta_individual[135, keebler]	0.012	0.048	-0.077	0.104	0.000	0.001	10419.0	3135.0	1.0
beta_individual[135, nabisco]	-0.009	0.051	-0.108	0.081	0.000	0.001	10504.0	2555.0	1.0
beta_individual[136, sunshine]	-0.002	0.049	-0.089	0.096	0.001	0.001	7867.0	2729.0	1.0
beta_individual[136, keebler]	-0.018	0.048	-0.112	0.071	0.001	0.001	7684.0	2845.0	1.0
beta_individual[136, nabisco]	0.020	0.050	-0.072	0.114	0.001	0.001	8252.0	2559.0	1.0

414 rows × 9 columns

What have we learned? We've imposed a negative slope on the price coefficient but given it a wide prior. We can see that the data is sufficient to have narrowed the likely range of the coefficient considerably.

az.plot_dist_comparison(idata_m4, var_names=["beta_price"]);

We have explicitly set a negative prior on price and recovered a parameter specification more in line with the basic theory of rational choice. The effect of price should have a negative impact on utility. The flexibility of priors here is key for incorporating theoretical knowledge about the process involved in choice. Priors are important for building a better picture of the decision making process and we'd be foolish to ignore their value in this setting.

fig, axs = plt.subplots(1, 2, figsize=(20, 10))
ax = axs[0]
counts = c_df.groupby("choice")["choiceId"].count()
labels = c_df.groupby("choice")["choiceId"].count().index
predicted_shares = idata_m4["posterior"]["p"].mean(dim=["chain", "draw", "obs"])
ci_lb = idata_m4["posterior"]["p"].quantile(0.025, dim=["chain", "draw", "obs"])
ci_ub = idata_m4["posterior"]["p"].quantile(0.975, dim=["chain", "draw", "obs"])
ax.scatter(ci_lb, labels, color="k", s=2)
ax.scatter(ci_ub, labels, color="k", s=2)
ax.scatter(
    counts / counts.sum(),
    labels,
    label="Observed Shares",
    color="red",
    s=100,
)
ax.scatter(
    predicted_shares,
    labels,
    label="Predicted Mean",
    color="green",
    s=100,
)
ax.hlines(
    labels,
    ci_lb,
    ci_ub,
    label="Predicted 95% Interval",
    color="black",
)
ax.legend()
ax.set_title("Observed V Predicted Shares")
az.plot_ppc(idata_m4, ax=axs[1])
axs[1].set_title("Posterior Predictive Checks")
ax.set_xlabel("Shares")
ax.set_ylabel("Crackers");

We can now also recover the differences among individuals estimated by the model for particular cracker choices. More precisely we'll plot how the individual specific contribution to the intercept drives preferences among the cracker choices.

idata_m4

arviz.InferenceData

• posterior

  <xarray.Dataset>
  Dimensions:          (chain: 4, draw: 1000, alts_intercepts: 3,
                        individuals: 136, obs: 3156, alts_probs: 4)
  Coordinates:
    * chain            (chain) int64 0 1 2 3
    * draw             (draw) int64 0 1 2 3 4 5 6 ... 993 994 995 996 997 998 999
    * alts_intercepts  (alts_intercepts) <U8 'sunshine' 'keebler' 'nabisco'
    * individuals      (individuals) int64 1 2 3 4 5 6 ... 131 132 133 134 135 136
    * obs              (obs) int64 0 1 2 3 4 5 6 ... 3150 3151 3152 3153 3154 3155
    * alts_probs       (alts_probs) <U8 'sunshine' 'keebler' 'nabisco' 'private'
  Data variables:
      alpha            (chain, draw, alts_intercepts) float64 -0.03604 ... 1.472
      beta_individual  (chain, draw, individuals, alts_intercepts) float64 -0.0...
      beta_feat        (chain, draw) float64 0.008284 0.00381 ... 0.06727 0.0183
      beta_disp        (chain, draw) float64 0.02731 0.02334 ... 0.003561 0.001434
      beta_price       (chain, draw) float64 -0.01967 -0.08042 ... -0.01557
      p                (chain, draw, obs, alts_probs) float64 0.05909 ... 0.07105
  Attributes:
      created_at:     2023-07-13T17:31:22.499390
      arviz_version:  0.15.1

  xarray.Dataset

  • Dimensions:
    • chain: 4
    • draw: 1000
    • alts_intercepts: 3
    • individuals: 136
    • obs: 3156
    • alts_probs: 4
  • Coordinates: (6)
    • chain
      (chain)
      int64
      0 1 2 3

      array([0, 1, 2, 3])

    • draw
      (draw)
      int64
      0 1 2 3 4 5 ... 995 996 997 998 999

      array([  0,   1,   2, ..., 997, 998, 999])

    • alts_intercepts
      (alts_intercepts)
      <U8
      'sunshine' 'keebler' 'nabisco'

      array(['sunshine', 'keebler', 'nabisco'], dtype='<U8')

    • individuals
      (individuals)
      int64
      1 2 3 4 5 6 ... 132 133 134 135 136

      array([  1,   2,   3,   4,   5,   6,   7,   8,   9,  10,  11,  12,  13,  14,
              15,  16,  17,  18,  19,  20,  21,  22,  23,  24,  25,  26,  27,  28,
              29,  30,  31,  32,  33,  34,  35,  36,  37,  38,  39,  40,  41,  42,
              43,  44,  45,  46,  47,  48,  49,  50,  51,  52,  53,  54,  55,  56,
              57,  58,  59,  60,  61,  62,  63,  64,  65,  66,  67,  68,  69,  70,
              71,  72,  73,  74,  75,  76,  77,  78,  79,  80,  81,  82,  83,  84,
              85,  86,  87,  88,  89,  90,  91,  92,  93,  94,  95,  96,  97,  98,
              99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112,
             113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126,
             127, 128, 129, 130, 131, 132, 133, 134, 135, 136])

    • obs
      (obs)
      int64
      0 1 2 3 4 ... 3152 3153 3154 3155

      array([   0,    1,    2, ..., 3153, 3154, 3155])

    • alts_probs
      (alts_probs)
      <U8
      'sunshine' 'keebler' ... 'private'

      array(['sunshine', 'keebler', 'nabisco', 'private'], dtype='<U8')

  • Data variables: (6)
    • alpha
      (chain, draw, alts_intercepts)
      float64
      -0.03604 2.145 ... 2.054 1.472

      array([[[-3.60354089e-02,  2.14480430e+00,  1.68506815e+00],
              [ 1.55227190e-01,  2.13629995e+00,  1.52518135e+00],
              [ 1.35868068e-02,  2.10626238e+00,  1.51838973e+00],
              ...,
              [ 7.54591026e-02,  2.10467698e+00,  1.54587101e+00],
              [ 4.04155687e-02,  2.04643441e+00,  1.50894838e+00],
              [ 9.31199468e-02,  2.10627525e+00,  1.56864008e+00]],
      
             [[-4.71215832e-02,  1.96048359e+00,  1.44187627e+00],
              [-4.12740981e-02,  2.05116312e+00,  1.48762707e+00],
              [-6.39949887e-02,  2.00165455e+00,  1.47207154e+00],
              ...,
              [ 7.53096155e-02,  2.05073796e+00,  1.53612143e+00],
              [-2.04418548e-01,  1.91268460e+00,  1.47868264e+00],
              [ 1.31145017e-03,  2.15266472e+00,  1.52350357e+00]],
      
             [[-1.67417178e-01,  1.93567031e+00,  1.36223077e+00],
              [-2.05434891e-03,  2.04217404e+00,  1.49425649e+00],
              [-9.62584618e-02,  1.97378688e+00,  1.46211638e+00],
              ...,
              [-1.78130585e-01,  1.89683295e+00,  1.42686467e+00],
              [-1.13593699e-01,  1.94827950e+00,  1.34688070e+00],
              [-1.00997139e-01,  2.03502095e+00,  1.42026456e+00]],
      
             [[-5.14730394e-02,  1.97567274e+00,  1.40884364e+00],
              [ 1.36100414e-02,  2.05271250e+00,  1.53461898e+00],
              [-1.28874270e-01,  1.94729367e+00,  1.47803578e+00],
              ...,
              [-1.89498299e-01,  1.92867822e+00,  1.42140180e+00],
              [-1.89138888e-01,  2.01432122e+00,  1.40835152e+00],
              [-1.60098982e-02,  2.05420574e+00,  1.47237351e+00]]])

    • beta_individual
      (chain, draw, individuals, alts_intercepts)
      float64
      -0.07106 0.01229 ... -0.02851

      array([[[[-7.10560114e-02,  1.22929332e-02, -1.46737437e-01],
               [ 8.52822212e-02,  3.32629274e-03, -8.20605473e-02],
               [ 4.78202552e-02, -2.18693778e-02, -3.68227935e-02],
               ...,
               [ 4.29781265e-02,  8.73897036e-02, -1.53910413e-02],
               [ 4.29146940e-03, -8.76279211e-03,  3.95867252e-02],
               [ 2.55974630e-02, -4.12636767e-02, -2.37613477e-02]],
      
              [[ 4.41778679e-02, -1.68296356e-02, -8.99795648e-02],
               [-1.29898100e-02, -7.51922913e-03, -3.36075829e-02],
               [ 5.36816645e-02, -5.50745672e-02,  8.18963024e-02],
               ...,
               [ 9.64938849e-03,  2.70293983e-02,  1.00558230e-02],
               [-8.83822929e-02,  8.21913566e-02,  3.26926018e-02],
               [-1.90090484e-02,  3.87284664e-02,  5.05434664e-02]],
      
              [[-5.77427761e-02, -1.70018539e-02,  4.47958404e-02],
               [ 5.26355345e-02, -2.71295228e-02, -2.47569760e-02],
               [-2.19562549e-02,  3.73185473e-02, -3.60250340e-02],
               ...,
      ...
               ...,
               [-9.61670811e-02,  8.47178812e-02,  6.74259652e-02],
               [ 7.27041214e-03, -7.87551611e-03, -4.26231534e-02],
               [-1.76720367e-02,  2.48984235e-02, -2.90015112e-02]],
      
              [[ 5.88698708e-02, -4.05063495e-02, -5.33633974e-03],
               [ 4.83901677e-02, -1.02670413e-01,  9.65700367e-03],
               [-8.92126797e-02,  5.18146129e-02,  4.22038044e-03],
               ...,
               [-1.34862027e-01,  2.07118333e-02, -7.52939259e-02],
               [ 4.88489501e-02, -4.17625557e-02, -4.55311955e-03],
               [-1.02991936e-01,  6.47643005e-02, -2.80800993e-02]],
      
              [[-1.04878612e-01, -7.22608778e-03, -6.50507719e-02],
               [-7.63817056e-02,  1.31133704e-02, -8.51631458e-02],
               [-7.41760347e-02, -7.44501815e-02,  4.95712014e-03],
               ...,
               [ 4.95819664e-02, -4.96280002e-02, -5.64554584e-02],
               [ 2.28594296e-02,  4.18671514e-02, -2.56275232e-02],
               [ 9.03650352e-02,  2.39682361e-02, -2.85149180e-02]]]])

    • beta_feat
      (chain, draw)
      float64
      0.008284 0.00381 ... 0.06727 0.0183

      array([[0.00828351, 0.00380971, 0.03502916, ..., 0.00189139, 0.01810111,
              0.01654725],
             [0.0039898 , 0.04145032, 0.01360652, ..., 0.01273682, 0.00126043,
              0.00874647],
             [0.01200379, 0.03659929, 0.00521525, ..., 0.00193413, 0.02779925,
              0.00242009],
             [0.02490679, 0.00953517, 0.03384639, ..., 0.02284003, 0.06727227,
              0.01829692]])

    • beta_disp
      (chain, draw)
      float64
      0.02731 0.02334 ... 0.001434

      array([[0.02731446, 0.02334214, 0.01891713, ..., 0.03182031, 0.00837427,
              0.03038024],
             [0.00222684, 0.11785599, 0.00092694, ..., 0.03481585, 0.02720738,
              0.06390792],
             [0.01221612, 0.01068079, 0.0197112 , ..., 0.04655466, 0.02036477,
              0.03293156],
             [0.00307374, 0.02925688, 0.00466769, ..., 0.001605  , 0.00356102,
              0.00143407]])

    • beta_price
      (chain, draw)
      float64
      -0.01967 -0.08042 ... -0.01557

      array([[-0.0196736 , -0.08041637, -0.00925359, ..., -0.03650781,
              -0.02930069, -0.04174801],
             [-0.005012  , -0.03464703, -0.02434125, ..., -0.05058218,
              -0.00947829, -0.00539878],
             [-0.01909274, -0.00570021, -0.00070166, ..., -0.03278035,
              -0.00531648, -0.00607888],
             [-0.01292809, -0.04037045, -0.00742446, ..., -0.00301366,
              -0.00907202, -0.01556731]])

    • p
      (chain, draw, obs, alts_probs)
      float64
      0.05909 0.5676 ... 0.2953 0.07105

      array([[[[0.05909284, 0.56755125, 0.30628985, 0.06706606],
               [0.06122941, 0.566625  , 0.30518898, 0.0669566 ],
               [0.0590135 , 0.56723547, 0.30672228, 0.06702874],
               ...,
               [0.06409828, 0.52997097, 0.33975595, 0.0661748 ],
               [0.0641317 , 0.52972592, 0.33993308, 0.0662093 ],
               [0.06375051, 0.53052877, 0.33964548, 0.06607524]],
      
              [[0.08267885, 0.55946863, 0.28450777, 0.07334475],
               [0.08782735, 0.55771401, 0.2813439 , 0.07311473],
               [0.08224624, 0.55833431, 0.28622341, 0.07319605],
               ...,
               [0.07280081, 0.55563083, 0.30221858, 0.06934977],
               [0.07296349, 0.55463785, 0.30289392, 0.06950474],
               [0.07120217, 0.55804393, 0.30182729, 0.06892661]],
      
              [[0.06459886, 0.54494668, 0.32231791, 0.06813655],
               [0.06605585, 0.54427133, 0.32162071, 0.06805211],
               [0.06455724, 0.54479713, 0.32252777, 0.06811786],
               ...,
      ...
               ...,
               [0.06306089, 0.54708639, 0.31202045, 0.07783227],
               [0.06306609, 0.54704905, 0.31204617, 0.07783868],
               [0.06300823, 0.54717338, 0.31200424, 0.07781415]],
      
              [[0.06677826, 0.54705771, 0.30940831, 0.07675572],
               [0.06730351, 0.54691423, 0.30904667, 0.07673559],
               [0.06673691, 0.54691737, 0.30960969, 0.07673603],
               ...,
               [0.05443504, 0.58259872, 0.2893381 , 0.07362815],
               [0.05444943, 0.58248841, 0.28941456, 0.0736476 ],
               [0.05429584, 0.5828512 , 0.28927973, 0.07357323]],
      
              [[0.06459119, 0.56362519, 0.29776045, 0.07402317],
               [0.06518053, 0.56354913, 0.29725715, 0.07401318],
               [0.06452365, 0.56338649, 0.29809804, 0.07399182],
               ...,
               [0.07533811, 0.55811196, 0.29541352, 0.07113641],
               [0.07537084, 0.55791999, 0.29554185, 0.07116732],
               [0.07501596, 0.55858765, 0.29534329, 0.07105311]]]])

  • Indexes: (6)
    • chain
      PandasIndex

      PandasIndex(Int64Index([0, 1, 2, 3], dtype='int64', name='chain'))

    • draw
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  990, 991, 992, 993, 994, 995, 996, 997, 998, 999],
                 dtype='int64', name='draw', length=1000))

    • alts_intercepts
      PandasIndex

      PandasIndex(Index(['sunshine', 'keebler', 'nabisco'], dtype='object', name='alts_intercepts'))

    • individuals
      PandasIndex

      PandasIndex(Int64Index([  1,   2,   3,   4,   5,   6,   7,   8,   9,  10,
                  ...
                  127, 128, 129, 130, 131, 132, 133, 134, 135, 136],
                 dtype='int64', name='individuals', length=136))

    • obs
      PandasIndex

      PandasIndex(Int64Index([   0,    1,    2,    3,    4,    5,    6,    7,    8,    9,
                  ...
                  3146, 3147, 3148, 3149, 3150, 3151, 3152, 3153, 3154, 3155],
                 dtype='int64', name='obs', length=3156))

    • alts_probs
      PandasIndex

      PandasIndex(Index(['sunshine', 'keebler', 'nabisco', 'private'], dtype='object', name='alts_probs'))

  • Attributes: (2)

    created_at :

    2023-07-13T17:31:22.499390

    arviz_version :

    0.15.1

  
  
• posterior_predictive

  <xarray.Dataset>
  Dimensions:  (chain: 4, draw: 1000, obs: 3156)
  Coordinates:
    * chain    (chain) int64 0 1 2 3
    * draw     (draw) int64 0 1 2 3 4 5 6 7 8 ... 992 993 994 995 996 997 998 999
    * obs      (obs) int64 0 1 2 3 4 5 6 7 ... 3149 3150 3151 3152 3153 3154 3155
  Data variables:
      y_cat    (chain, draw, obs) int64 2 0 1 2 2 1 2 2 1 1 ... 2 1 2 2 1 1 2 2 1
  Attributes:
      created_at:                 2023-07-13T17:31:45.522950
      arviz_version:              0.15.1
      inference_library:          pymc
      inference_library_version:  5.3.0

  xarray.Dataset

  • Dimensions:
    • chain: 4
    • draw: 1000
    • obs: 3156
  • Coordinates: (3)
    • chain
      (chain)
      int64
      0 1 2 3

      array([0, 1, 2, 3])

    • draw
      (draw)
      int64
      0 1 2 3 4 5 ... 995 996 997 998 999

      array([  0,   1,   2, ..., 997, 998, 999])

    • obs
      (obs)
      int64
      0 1 2 3 4 ... 3152 3153 3154 3155

      array([   0,    1,    2, ..., 3153, 3154, 3155])

  • Data variables: (1)
    • y_cat
      (chain, draw, obs)
      int64
      2 0 1 2 2 1 2 2 ... 1 2 2 1 1 2 2 1

      array([[[2, 0, 1, ..., 0, 2, 2],
              [2, 1, 2, ..., 0, 1, 1],
              [0, 2, 1, ..., 1, 1, 1],
              ...,
              [2, 1, 2, ..., 1, 2, 1],
              [3, 1, 1, ..., 3, 2, 0],
              [1, 1, 2, ..., 2, 3, 2]],
      
             [[3, 1, 1, ..., 0, 1, 3],
              [1, 1, 1, ..., 1, 1, 2],
              [2, 3, 2, ..., 1, 1, 1],
              ...,
              [1, 1, 1, ..., 2, 1, 1],
              [1, 2, 1, ..., 3, 1, 1],
              [1, 1, 1, ..., 1, 1, 1]],
      
             [[1, 2, 3, ..., 1, 2, 1],
              [2, 1, 1, ..., 0, 1, 1],
              [1, 1, 0, ..., 1, 1, 1],
              ...,
              [1, 1, 3, ..., 2, 1, 3],
              [2, 1, 1, ..., 1, 1, 1],
              [2, 1, 2, ..., 0, 1, 1]],
      
             [[1, 1, 2, ..., 1, 3, 1],
              [1, 1, 3, ..., 1, 3, 2],
              [2, 0, 2, ..., 2, 1, 3],
              ...,
              [1, 1, 2, ..., 2, 1, 2],
              [1, 3, 0, ..., 1, 1, 1],
              [0, 3, 1, ..., 2, 2, 1]]])

  • Indexes: (3)
    • chain
      PandasIndex

      PandasIndex(Int64Index([0, 1, 2, 3], dtype='int64', name='chain'))

    • draw
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  990, 991, 992, 993, 994, 995, 996, 997, 998, 999],
                 dtype='int64', name='draw', length=1000))

    • obs
      PandasIndex

      PandasIndex(Int64Index([   0,    1,    2,    3,    4,    5,    6,    7,    8,    9,
                  ...
                  3146, 3147, 3148, 3149, 3150, 3151, 3152, 3153, 3154, 3155],
                 dtype='int64', name='obs', length=3156))

  • Attributes: (4)

    created_at :

    2023-07-13T17:31:45.522950

    arviz_version :

    0.15.1

    inference_library :

    pymc

    inference_library_version :

    5.3.0

  
  
• log_likelihood

  <xarray.Dataset>
  Dimensions:      (chain: 4, draw: 1000, y_cat_dim_0: 1, y_cat_dim_1: 3156)
  Coordinates:
    * chain        (chain) int64 0 1 2 3
    * draw         (draw) int64 0 1 2 3 4 5 6 7 ... 993 994 995 996 997 998 999
    * y_cat_dim_0  (y_cat_dim_0) int64 0
    * y_cat_dim_1  (y_cat_dim_1) int64 0 1 2 3 4 5 ... 3151 3152 3153 3154 3155
  Data variables:
      y_cat        (chain, draw, y_cat_dim_0, y_cat_dim_1) float64 -0.5664 ... ...
  Attributes:
      created_at:     2023-07-13T17:31:22.503457
      arviz_version:  0.15.1

  xarray.Dataset

  • Dimensions:
    • chain: 4
    • draw: 1000
    • y_cat_dim_0: 1
    • y_cat_dim_1: 3156
  • Coordinates: (4)
    • chain
      (chain)
      int64
      0 1 2 3

      array([0, 1, 2, 3])

    • draw
      (draw)
      int64
      0 1 2 3 4 5 ... 995 996 997 998 999

      array([  0,   1,   2, ..., 997, 998, 999])

    • y_cat_dim_0
      (y_cat_dim_0)
      int64
      0

      array([0])

    • y_cat_dim_1
      (y_cat_dim_1)
      int64
      0 1 2 3 4 ... 3152 3153 3154 3155

      array([   0,    1,    2, ..., 3153, 3154, 3155])

  • Data variables: (1)
    • y_cat
      (chain, draw, y_cat_dim_0, y_cat_dim_1)
      float64
      -0.5664 -2.704 ... -1.219 -1.22

      array([[[[-0.56642423, -2.70371058, -0.56698076, ..., -1.07952771,
                -1.07900651, -1.0798529 ]],
      
              [[-0.58076783, -2.61572544, -0.58279738, ..., -1.19660474,
                -1.19437265, -1.19790031]],
      
              [[-0.60706733, -2.68748151, -0.60734179, ..., -1.14010482,
                -1.1398558 , -1.14024032]],
      
              ...,
      
              [[-0.59519723, -2.73805388, -0.59622746, ..., -1.17820235,
                -1.177196  , -1.17878314]],
      
              [[-0.60855422, -2.57035298, -0.60937226, ..., -1.09977667,
                -1.0990142 , -1.10016394]],
      
              [[-0.61709514, -2.7461461 , -0.61835305, ..., -1.12361282,
                -1.12251009, -1.12417094]]],
      
      ...
      
             [[[-0.58845121, -2.62863572, -0.5888036 , ..., -1.2043197 ,
                -1.20396418, -1.2045135 ]],
      
              [[-0.62514198, -2.58764011, -0.62631403, ..., -1.07889076,
                -1.07784404, -1.07946088]],
      
              [[-0.62164258, -2.59630761, -0.62186575, ..., -1.1671637 ,
                -1.16695938, -1.16729258]],
      
              ...,
      
              [[-0.6226461 , -2.55306248, -0.62273306, ..., -1.16468655,
                -1.16460412, -1.16473851]],
      
              [[-0.60320098, -2.56738966, -0.60345755, ..., -1.24015939,
                -1.23989514, -1.24036115]],
      
              [[-0.57336581, -2.60351204, -0.5737894 , ..., -1.21937916,
                -1.21894481, -1.21961691]]]])

  • Indexes: (4)
    • chain
      PandasIndex

      PandasIndex(Int64Index([0, 1, 2, 3], dtype='int64', name='chain'))

    • draw
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  990, 991, 992, 993, 994, 995, 996, 997, 998, 999],
                 dtype='int64', name='draw', length=1000))

    • y_cat_dim_0
      PandasIndex

      PandasIndex(Int64Index([0], dtype='int64', name='y_cat_dim_0'))

    • y_cat_dim_1
      PandasIndex

      PandasIndex(Int64Index([   0,    1,    2,    3,    4,    5,    6,    7,    8,    9,
                  ...
                  3146, 3147, 3148, 3149, 3150, 3151, 3152, 3153, 3154, 3155],
                 dtype='int64', name='y_cat_dim_1', length=3156))

  • Attributes: (2)

    created_at :

    2023-07-13T17:31:22.503457

    arviz_version :

    0.15.1

  
  
• sample_stats

  <xarray.Dataset>
  Dimensions:          (chain: 4, draw: 1000)
  Coordinates:
    * chain            (chain) int64 0 1 2 3
    * draw             (draw) int64 0 1 2 3 4 5 6 ... 993 994 995 996 997 998 999
  Data variables:
      acceptance_rate  (chain, draw) float64 0.7536 1.0 0.832 ... 0.5635 0.7482
      step_size        (chain, draw) float64 0.3084 0.3084 ... 0.3254 0.3254
      diverging        (chain, draw) bool False False False ... False False False
      energy           (chain, draw) float64 2.955e+03 2.943e+03 ... 2.922e+03
      n_steps          (chain, draw) int64 15 15 15 15 15 15 ... 15 15 15 15 15 15
      tree_depth       (chain, draw) int64 4 4 4 4 4 4 4 4 4 ... 4 4 4 4 4 4 4 4 4
      lp               (chain, draw) float64 2.737e+03 2.724e+03 ... 2.732e+03
  Attributes:
      created_at:     2023-07-13T17:31:22.502328
      arviz_version:  0.15.1

  xarray.Dataset

  • Dimensions:
    • chain: 4
    • draw: 1000
  • Coordinates: (2)
    • chain
      (chain)
      int64
      0 1 2 3

      array([0, 1, 2, 3])

    • draw
      (draw)
      int64
      0 1 2 3 4 5 ... 995 996 997 998 999

      array([  0,   1,   2, ..., 997, 998, 999])

  • Data variables: (7)
    • acceptance_rate
      (chain, draw)
      float64
      0.7536 1.0 0.832 ... 0.5635 0.7482

      array([[0.75359483, 1.        , 0.83195885, ..., 0.76011887, 0.99787025,
              0.86046751],
             [0.95927329, 0.90971071, 0.81827508, ..., 0.95107498, 0.64816401,
              0.99841539],
             [0.84699449, 0.9766036 , 0.83339206, ..., 0.88261802, 0.96639012,
              0.55951519],
             [0.71184684, 0.52649917, 0.90061833, ..., 0.9115007 , 0.56345889,
              0.74819169]])

    • step_size
      (chain, draw)
      float64
      0.3084 0.3084 ... 0.3254 0.3254

      array([[0.30835069, 0.30835069, 0.30835069, ..., 0.30835069, 0.30835069,
              0.30835069],
             [0.27911576, 0.27911576, 0.27911576, ..., 0.27911576, 0.27911576,
              0.27911576],
             [0.29232966, 0.29232966, 0.29232966, ..., 0.29232966, 0.29232966,
              0.29232966],
             [0.32540986, 0.32540986, 0.32540986, ..., 0.32540986, 0.32540986,
              0.32540986]])

    • diverging
      (chain, draw)
      bool
      False False False ... False False

      array([[False, False, False, ..., False, False, False],
             [False, False, False, ..., False, False, False],
             [False, False, False, ..., False, False, False],
             [False, False, False, ..., False, False, False]])

    • energy
      (chain, draw)
      float64
      2.955e+03 2.943e+03 ... 2.922e+03

      array([[2955.32125472, 2942.63944741, 2937.38368085, ..., 2958.80891917,
              2910.46509685, 2951.65754222],
             [2921.30393409, 2918.98886851, 2932.15951712, ..., 2877.89864067,
              2938.04224536, 2930.1388167 ],
             [2922.73108687, 2913.64412748, 2918.11622689, ..., 2938.22447495,
              2912.23541398, 2948.76780202],
             [2934.70547972, 2940.96383695, 2902.30383696, ..., 2887.25745514,
              2910.3572611 , 2921.75197175]])

    • n_steps
      (chain, draw)
      int64
      15 15 15 15 15 ... 15 15 15 15 15

      array([[15, 15, 15, ..., 15, 15, 15],
             [15, 15, 15, ..., 15, 15, 15],
             [15, 15, 15, ..., 15, 15, 15],
             [15, 15, 15, ..., 15, 15, 15]])

    • tree_depth
      (chain, draw)
      int64
      4 4 4 4 4 4 4 4 ... 4 4 4 4 4 4 4 4

      array([[4, 4, 4, ..., 4, 4, 4],
             [4, 4, 4, ..., 4, 4, 4],
             [4, 4, 4, ..., 4, 4, 4],
             [4, 4, 4, ..., 4, 4, 4]])

    • lp
      (chain, draw)
      float64
      2.737e+03 2.724e+03 ... 2.732e+03

      array([[2736.77350266, 2724.31496193, 2706.31508118, ..., 2730.37254665,
              2727.86405196, 2746.01564599],
             [2712.01791995, 2722.61890377, 2729.85651359, ..., 2703.10501159,
              2725.93956613, 2718.5276056 ],
             [2714.00643247, 2711.85104302, 2731.73364302, ..., 2722.14845037,
              2704.75333532, 2734.63229096],
             [2705.05439772, 2707.31530869, 2699.7048273 , ..., 2694.93775834,
              2714.42031119, 2732.4017896 ]])

  • Indexes: (2)
    • chain
      PandasIndex

      PandasIndex(Int64Index([0, 1, 2, 3], dtype='int64', name='chain'))

    • draw
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  990, 991, 992, 993, 994, 995, 996, 997, 998, 999],
                 dtype='int64', name='draw', length=1000))

  • Attributes: (2)

    created_at :

    2023-07-13T17:31:22.502328

    arviz_version :

    0.15.1

  
  
• prior

  <xarray.Dataset>
  Dimensions:          (chain: 1, draw: 500, individuals: 136,
                        alts_intercepts: 3, obs: 3156, alts_probs: 4)
  Coordinates:
    * chain            (chain) int64 0
    * draw             (draw) int64 0 1 2 3 4 5 6 ... 493 494 495 496 497 498 499
    * individuals      (individuals) int64 1 2 3 4 5 6 ... 131 132 133 134 135 136
    * alts_intercepts  (alts_intercepts) <U8 'sunshine' 'keebler' 'nabisco'
    * obs              (obs) int64 0 1 2 3 4 5 6 ... 3150 3151 3152 3153 3154 3155
    * alts_probs       (alts_probs) <U8 'sunshine' 'keebler' 'nabisco' 'private'
  Data variables:
      beta_individual  (chain, draw, individuals, alts_intercepts) float64 -0.0...
      p                (chain, draw, obs, alts_probs) float64 0.4208 ... 0.4889
      beta_feat        (chain, draw) float64 1.523 0.1104 0.4152 ... 0.6374 0.1037
      beta_price       (chain, draw) float64 -0.06909 -1.517 ... -0.136 -1.223
      beta_disp        (chain, draw) float64 1.26 0.4421 2.428 ... 1.101 1.011
      alpha            (chain, draw, alts_intercepts) float64 0.8138 ... 0.5752
  Attributes:
      created_at:                 2023-07-13T17:31:04.376498
      arviz_version:              0.15.1
      inference_library:          pymc
      inference_library_version:  5.3.0

  xarray.Dataset

  • Dimensions:
    • chain: 1
    • draw: 500
    • individuals: 136
    • alts_intercepts: 3
    • obs: 3156
    • alts_probs: 4
  • Coordinates: (6)
    • chain
      (chain)
      int64
      0

      array([0])

    • draw
      (draw)
      int64
      0 1 2 3 4 5 ... 495 496 497 498 499

      array([  0,   1,   2, ..., 497, 498, 499])

    • individuals
      (individuals)
      int64
      1 2 3 4 5 6 ... 132 133 134 135 136

      array([  1,   2,   3,   4,   5,   6,   7,   8,   9,  10,  11,  12,  13,  14,
              15,  16,  17,  18,  19,  20,  21,  22,  23,  24,  25,  26,  27,  28,
              29,  30,  31,  32,  33,  34,  35,  36,  37,  38,  39,  40,  41,  42,
              43,  44,  45,  46,  47,  48,  49,  50,  51,  52,  53,  54,  55,  56,
              57,  58,  59,  60,  61,  62,  63,  64,  65,  66,  67,  68,  69,  70,
              71,  72,  73,  74,  75,  76,  77,  78,  79,  80,  81,  82,  83,  84,
              85,  86,  87,  88,  89,  90,  91,  92,  93,  94,  95,  96,  97,  98,
              99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112,
             113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126,
             127, 128, 129, 130, 131, 132, 133, 134, 135, 136])

    • alts_intercepts
      (alts_intercepts)
      <U8
      'sunshine' 'keebler' 'nabisco'

      array(['sunshine', 'keebler', 'nabisco'], dtype='<U8')

    • obs
      (obs)
      int64
      0 1 2 3 4 ... 3152 3153 3154 3155

      array([   0,    1,    2, ..., 3153, 3154, 3155])

    • alts_probs
      (alts_probs)
      <U8
      'sunshine' 'keebler' ... 'private'

      array(['sunshine', 'keebler', 'nabisco', 'private'], dtype='<U8')

  • Data variables: (6)
    • beta_individual
      (chain, draw, individuals, alts_intercepts)
      float64
      -0.02606 0.08851 ... -0.01381

      array([[[[-0.0260648 ,  0.08850873,  0.00852248],
               [-0.00581379,  0.05465033,  0.00524874],
               [-0.07100896,  0.07007681,  0.08090251],
               ...,
               [-0.03483249, -0.0420326 , -0.08291526],
               [-0.03591725, -0.01692773, -0.08138898],
               [-0.05205832,  0.00754408, -0.0481465 ]],
      
              [[-0.02672593,  0.03912129, -0.05592089],
               [-0.00924045,  0.01144507, -0.0331788 ],
               [ 0.08603022,  0.00779349, -0.00632146],
               ...,
               [ 0.01522275,  0.0079651 ,  0.03764743],
               [-0.00588655, -0.02029812,  0.04883813],
               [-0.0377978 ,  0.02608773, -0.08994365]],
      
              [[ 0.06693098, -0.05578379, -0.05750472],
               [ 0.03448372, -0.00401051,  0.03395676],
               [-0.01498093,  0.01666808, -0.03924826],
               ...,
      ...
               ...,
               [-0.03253887, -0.03567215, -0.03627986],
               [-0.02976413, -0.02343611,  0.02906421],
               [ 0.01162583, -0.03910649,  0.05774165]],
      
              [[ 0.00350196, -0.03307139,  0.0256701 ],
               [-0.01220089, -0.02621338,  0.04797904],
               [-0.05185757,  0.01165223, -0.06179474],
               ...,
               [-0.05010166,  0.02363455, -0.06245668],
               [ 0.04367377, -0.07000926, -0.01361819],
               [-0.01647258,  0.07144693,  0.0515059 ]],
      
              [[ 0.02672852,  0.06559873, -0.02723333],
               [ 0.02357888, -0.01546331,  0.0537435 ],
               [ 0.05950198, -0.02196791, -0.01205404],
               ...,
               [ 0.01193274,  0.02561176,  0.15508737],
               [ 0.03234659, -0.11740604,  0.05785968],
               [ 0.05167683, -0.03551028, -0.01380554]]]])

    • p
      (chain, draw, obs, alts_probs)
      float64
      0.4208 0.1497 ... 0.1903 0.4889

      array([[[[0.42081506, 0.14974991, 0.22446748, 0.20496755],
               [0.72674377, 0.07084019, 0.10545477, 0.09696127],
               [0.41948812, 0.1496908 , 0.22593442, 0.20488665],
               ...,
               [0.42537766, 0.14347505, 0.21702704, 0.21412026],
               [0.42558822, 0.14305107, 0.21713447, 0.21422625],
               [0.42107373, 0.14529852, 0.21872497, 0.21490278]],
      
              [[0.59436071, 0.09130586, 0.01061746, 0.30371598],
               [0.83012956, 0.03837778, 0.00383444, 0.12765823],
               [0.57860431, 0.09444741, 0.01278238, 0.3141659 ],
               ...,
               [0.56424505, 0.08917524, 0.0072723 , 0.33930742],
               [0.56794545, 0.08320192, 0.00731999, 0.34153264],
               [0.47745274, 0.12450178, 0.00913005, 0.38891543]],
      
              [[0.09647545, 0.11894673, 0.59944254, 0.18513528],
               [0.57276439, 0.05692229, 0.2817163 , 0.08859701],
               [0.09480646, 0.11773885, 0.60419943, 0.18325527],
               ...,
      ...
               ...,
               [0.24422187, 0.11121754, 0.12834168, 0.5162189 ],
               [0.24516662, 0.10777939, 0.12883816, 0.51821583],
               [0.21552075, 0.12386605, 0.13605256, 0.52456064]],
      
              [[0.21361527, 0.2815044 , 0.33054011, 0.17434022],
               [0.46781999, 0.19159347, 0.22192961, 0.11865693],
               [0.21174401, 0.28056003, 0.3339406 , 0.17375536],
               ...,
               [0.20337487, 0.30436705, 0.32060884, 0.17164923],
               [0.2037951 , 0.3029297 , 0.3212713 , 0.1720039 ],
               [0.19740661, 0.30899159, 0.32239731, 0.17120448]],
      
              [[0.1195044 , 0.21263683, 0.22709628, 0.4407625 ],
               [0.41486254, 0.145634  , 0.13762721, 0.30187625],
               [0.11114688, 0.20768441, 0.25067179, 0.43049692],
               ...,
               [0.12279298, 0.19738227, 0.18060772, 0.49921704],
               [0.12424813, 0.18787088, 0.182748  , 0.50513299],
               [0.09415265, 0.22661895, 0.19034145, 0.48888695]]]])

    • beta_feat
      (chain, draw)
      float64
      1.523 0.1104 ... 0.6374 0.1037

      array([[1.52277970e+00, 1.10376852e-01, 4.15216147e-01, 1.91779124e+00,
              8.53840561e-01, 3.26806946e+00, 8.02014014e-01, 1.30652478e-01,
              2.05279697e-02, 3.10785232e-02, 2.01703404e-01, 1.02112770e+00,
              5.93153181e-01, 2.83135196e-01, 9.18968471e-01, 2.88570776e+00,
              8.88156866e-01, 2.28445905e-01, 1.04508781e-02, 1.05582839e+00,
              1.97050223e-01, 2.19883645e-02, 4.26422429e-01, 5.37714586e-02,
              1.88193897e-01, 5.99217590e-01, 9.15832816e-01, 7.94140163e-01,
              4.21412363e-01, 6.95548417e-02, 7.42568283e-01, 2.01936872e-02,
              7.14190254e-01, 3.22456941e-01, 3.85971210e-01, 1.34720718e+00,
              2.42904229e-01, 1.04063344e-01, 1.63539107e+00, 4.90321914e-01,
              3.15075451e-01, 2.74678746e-01, 2.07735077e+00, 1.00015686e+00,
              2.24149416e-01, 5.12304645e-01, 1.16162755e+00, 5.44136467e-02,
              2.14552482e-01, 7.44560689e-01, 4.59272079e-01, 4.55684728e-01,
              8.28521240e-01, 6.10943635e-01, 9.40711602e-01, 2.45650332e-01,
              1.78103211e+00, 4.14209268e-01, 7.82337778e-01, 1.51105196e+00,
              1.25749959e-01, 3.20352751e-01, 1.39802260e+00, 1.68168952e-01,
              6.71515898e-01, 9.28549039e-01, 5.59694515e-01, 6.54399026e-01,
              9.25131841e-01, 3.67317864e-01, 6.69006421e-01, 8.09089135e-01,
              1.56480579e-01, 1.41280032e+00, 1.28263661e+00, 3.21774509e-01,
              1.16519425e+00, 3.97200945e-01, 2.93401493e-01, 1.16709058e+00,
      ...
              2.20604466e-02, 1.93595460e-01, 4.77446620e-01, 7.89142182e-01,
              8.91032600e-01, 2.67405810e+00, 1.78674817e+00, 1.37991940e+00,
              2.30987749e+00, 6.95313287e-02, 1.15549460e+00, 1.11900127e+00,
              4.44427163e-01, 1.67956515e+00, 9.92137017e-02, 1.91144231e+00,
              1.26143631e+00, 4.05749383e-01, 1.50718312e-01, 5.82155052e-01,
              1.14349559e+00, 2.06503187e-02, 3.58222099e-01, 1.05486406e+00,
              3.43686628e-01, 4.83335524e-01, 4.40221296e-01, 9.52726962e-03,
              2.19965513e+00, 6.77667386e-01, 6.94527769e-01, 1.00436308e-01,
              7.87098974e-01, 1.36502824e+00, 1.50207841e-01, 1.26760286e+00,
              4.50146151e-01, 1.65757579e+00, 1.23711986e+00, 1.42145645e+00,
              6.79664041e-01, 8.52125542e-01, 8.35810783e-01, 2.23689100e+00,
              1.60108972e+00, 2.16900773e-01, 6.20468172e-01, 8.71621915e-01,
              6.24334523e-02, 1.71408596e+00, 1.88558818e+00, 6.31363848e-01,
              1.11960331e-01, 3.35765322e-01, 1.92281993e+00, 1.98111127e+00,
              1.66059767e+00, 2.56998475e+00, 1.15173428e+00, 5.18468680e-02,
              1.73850173e+00, 9.09325000e-01, 8.83063893e-01, 1.46542719e+00,
              1.48154177e+00, 1.49815590e+00, 1.09485053e+00, 1.33989673e-01,
              2.02316919e+00, 1.71145920e+00, 6.84550548e-01, 1.87228101e+00,
              5.42470820e-01, 1.69610828e+00, 8.70557039e-02, 2.83066957e-01,
              2.40616361e-01, 6.62950388e-01, 6.37414978e-01, 1.03671519e-01]])

    • beta_price
      (chain, draw)
      float64
      -0.06909 -1.517 ... -0.136 -1.223

      array([[-6.90868271e-02, -1.51739420e+00, -1.81109278e-01,
              -4.35710277e-01, -7.44308084e-01, -2.55442722e-01,
              -6.05214623e-01, -7.98113040e-01, -1.01890915e+00,
              -7.23328232e-01, -6.46299171e-01, -1.23558706e+00,
              -4.80830034e-01, -1.28781859e+00, -6.56814153e-02,
              -4.25230873e-02, -4.72084420e-01, -1.18924014e+00,
              -3.17210539e-02, -1.14714043e+00, -2.07654062e+00,
              -1.19083447e+00, -4.64586465e-01, -1.00101447e+00,
              -3.27653308e-01, -1.16939220e+00, -3.02689553e-01,
              -7.91376164e-01, -5.08120944e-01, -4.92009205e-01,
              -5.14427066e-02, -3.38129100e-01, -4.78531065e-01,
              -8.18380845e-01, -6.01410224e-01, -1.10830283e-01,
              -1.91201518e-01, -8.88865214e-02, -2.86801858e-01,
              -1.45583406e+00, -4.13375094e-01, -8.53371775e-02,
              -4.34363611e-01, -5.20435635e-01, -2.71488164e-01,
              -8.98123685e-02, -5.62996124e-01, -1.23477146e+00,
              -1.44979003e+00, -1.18501178e+00, -5.32043101e-01,
              -1.01009891e+00, -7.65064699e-01, -8.11513423e-01,
              -1.51733617e+00, -1.02114963e+00, -5.21002178e-01,
              -1.29767787e+00, -1.47340731e+00, -3.47159934e-01,
      ...
              -3.64368621e-01, -2.25611254e-01, -6.25295417e-01,
              -2.15768214e+00, -6.30862635e-01, -9.69440094e-01,
              -2.22608846e-01, -2.60408987e-01, -1.30242365e+00,
              -9.95998549e-01, -1.23360336e+00, -1.87688418e+00,
              -3.36962535e-01, -1.88141624e+00, -8.56426230e-01,
              -1.92428841e-01, -2.33695365e-02, -2.99540395e-02,
              -3.66728169e-01, -5.67784090e-01, -1.86785197e+00,
              -5.07515730e-01, -6.60707072e-01, -4.29145872e-01,
              -7.84756503e-01, -1.39512271e+00, -2.05200969e-01,
              -1.57640683e+00, -6.23117822e-03, -4.01585747e-01,
              -7.72178837e-01, -1.08466376e+00, -3.75540926e-01,
              -6.56860576e-01, -2.57213299e-01, -8.11404062e-01,
              -1.41857758e+00, -2.15809349e+00, -5.88865386e-01,
              -1.00923950e+00, -5.86894393e-01, -1.94522037e-01,
              -1.13192274e+00, -3.33608429e-02, -6.69482622e-01,
              -1.01172915e+00, -1.62971571e+00, -9.09129178e-01,
              -1.22400917e+00, -2.73365219e-01, -1.27927120e+00,
              -6.64463511e-01, -5.14741081e-02, -1.21260834e+00,
              -3.31759776e-01, -3.44405499e-01, -7.05250380e-01,
              -1.35954564e-01, -1.22336417e+00]])

    • beta_disp
      (chain, draw)
      float64
      1.26 0.4421 2.428 ... 1.101 1.011

      array([[1.26037718e+00, 4.42134701e-01, 2.42761985e+00, 3.14420820e-01,
              1.62612574e+00, 1.25466868e-02, 1.05991409e-02, 6.27694404e-01,
              1.02277707e-01, 1.43251238e+00, 1.09436099e+00, 4.16370774e-01,
              2.58263757e-01, 1.91891937e+00, 7.61765695e-02, 6.48994269e-01,
              4.14140002e-01, 3.04046086e-01, 2.20272656e+00, 1.52882762e-01,
              8.15552659e-02, 1.21534102e+00, 1.17137057e-01, 6.10625274e-01,
              7.67532095e-01, 4.47971096e-01, 1.95071661e+00, 1.40224266e-01,
              5.40504437e-01, 7.75842701e-01, 3.70507921e-01, 8.43020450e-01,
              3.12548639e-01, 4.52288949e-01, 1.01541822e+00, 8.31812346e-02,
              1.81167958e+00, 2.65550311e+00, 8.73617322e-01, 1.41807704e+00,
              3.49109702e-01, 2.78127957e-03, 1.40531353e+00, 3.58478895e-01,
              3.03886740e-01, 5.94634821e-01, 1.13387196e+00, 7.77946261e-02,
              1.61483033e+00, 1.05339127e+00, 1.26102744e+00, 4.46205395e-01,
              1.74250805e+00, 4.11218186e-01, 1.03216305e+00, 2.85928171e-01,
              4.75477950e-01, 8.49116931e-01, 9.85868050e-01, 1.26266718e-01,
              2.88261565e-01, 7.88027561e-01, 5.42623175e-01, 5.41211774e-01,
              1.31612110e-01, 1.87248978e-01, 1.35460499e+00, 5.83234340e-01,
              9.17008902e-01, 1.98625324e-01, 5.46300902e-01, 3.12733959e-01,
              1.42885039e+00, 6.65390082e-01, 5.91891444e-01, 3.12250494e-01,
              1.51192553e+00, 6.86042097e-01, 1.30821562e+00, 1.29743903e+00,
      ...
              1.09873757e+00, 8.82790835e-01, 1.46250007e+00, 2.19279859e-01,
              4.83880593e-01, 5.52758029e-01, 7.00255171e-01, 7.99754936e-01,
              3.77283656e-02, 2.92437236e-01, 1.41139562e+00, 1.06273977e+00,
              7.52236689e-01, 8.26425728e-01, 1.15370811e+00, 3.28251844e+00,
              1.00087087e+00, 1.98644276e+00, 1.09166872e+00, 1.06589841e+00,
              1.74253005e+00, 1.64665581e+00, 6.56049979e-01, 4.97643305e-01,
              2.22307231e-01, 9.17804794e-01, 9.30331155e-02, 1.23389841e+00,
              8.28416535e-02, 2.78115378e-01, 8.70394443e-01, 2.04911325e-01,
              1.11822964e-02, 1.17028122e+00, 1.00347201e+00, 6.49705516e-02,
              1.37048394e+00, 6.09144992e-01, 1.31148209e+00, 9.08774978e-01,
              1.60451807e-01, 1.56038648e+00, 8.64385481e-01, 3.45272090e-01,
              5.98057241e-01, 8.40439427e-02, 3.76154979e-01, 9.72542903e-01,
              1.68357644e+00, 5.15356366e-01, 1.10842421e+00, 1.68783558e+00,
              2.15298808e-01, 1.01109185e+00, 2.23054731e-01, 5.27478741e-01,
              2.97035190e-02, 9.74809573e-01, 1.74878169e-01, 5.47054367e-01,
              2.65165913e+00, 5.47471302e-01, 3.27717434e-01, 1.73398095e+00,
              7.86364980e-01, 1.88362390e+00, 4.04698218e-01, 1.91804326e+00,
              8.42644166e-01, 2.95664822e-01, 9.99306860e-01, 1.66392353e+00,
              3.62520797e-01, 7.44921636e-01, 1.12069118e+00, 4.37878902e-01,
              1.65330373e+00, 1.27383969e+00, 1.10070203e+00, 1.01140101e+00]])

    • alpha
      (chain, draw, alts_intercepts)
      float64
      0.8138 -0.3271 ... 0.5389 0.5752

      array([[[ 0.81380253, -0.32708916,  0.15075267],
              [ 2.20033959,  0.41296032, -1.79545215],
              [-0.53943099, -0.18921816,  1.41171621],
              ...,
              [ 0.0086433 , -0.67079339, -0.53980345],
              [ 0.33426094,  0.66040123,  0.74864425],
              [-0.12075098,  0.53894772,  0.57523174]]])

  • Indexes: (6)
    • chain
      PandasIndex

      PandasIndex(Int64Index([0], dtype='int64', name='chain'))

    • draw
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  490, 491, 492, 493, 494, 495, 496, 497, 498, 499],
                 dtype='int64', name='draw', length=500))

    • individuals
      PandasIndex

      PandasIndex(Int64Index([  1,   2,   3,   4,   5,   6,   7,   8,   9,  10,
                  ...
                  127, 128, 129, 130, 131, 132, 133, 134, 135, 136],
                 dtype='int64', name='individuals', length=136))

    • alts_intercepts
      PandasIndex

      PandasIndex(Index(['sunshine', 'keebler', 'nabisco'], dtype='object', name='alts_intercepts'))

    • obs
      PandasIndex

      PandasIndex(Int64Index([   0,    1,    2,    3,    4,    5,    6,    7,    8,    9,
                  ...
                  3146, 3147, 3148, 3149, 3150, 3151, 3152, 3153, 3154, 3155],
                 dtype='int64', name='obs', length=3156))

    • alts_probs
      PandasIndex

      PandasIndex(Index(['sunshine', 'keebler', 'nabisco', 'private'], dtype='object', name='alts_probs'))

  • Attributes: (4)

    created_at :

    2023-07-13T17:31:04.376498

    arviz_version :

    0.15.1

    inference_library :

    pymc

    inference_library_version :

    5.3.0

  
  
• prior_predictive

  <xarray.Dataset>
  Dimensions:  (chain: 1, draw: 500, obs: 3156)
  Coordinates:
    * chain    (chain) int64 0
    * draw     (draw) int64 0 1 2 3 4 5 6 7 8 ... 492 493 494 495 496 497 498 499
    * obs      (obs) int64 0 1 2 3 4 5 6 7 ... 3149 3150 3151 3152 3153 3154 3155
  Data variables:
      y_cat    (chain, draw, obs) int64 2 1 3 1 0 2 3 1 2 2 ... 0 0 1 1 2 3 3 0 2
  Attributes:
      created_at:                 2023-07-13T17:31:04.377859
      arviz_version:              0.15.1
      inference_library:          pymc
      inference_library_version:  5.3.0

  xarray.Dataset

  • Dimensions:
    • chain: 1
    • draw: 500
    • obs: 3156
  • Coordinates: (3)
    • chain
      (chain)
      int64
      0

      array([0])

    • draw
      (draw)
      int64
      0 1 2 3 4 5 ... 495 496 497 498 499

      array([  0,   1,   2, ..., 497, 498, 499])

    • obs
      (obs)
      int64
      0 1 2 3 4 ... 3152 3153 3154 3155

      array([   0,    1,    2, ..., 3153, 3154, 3155])

  • Data variables: (1)
    • y_cat
      (chain, draw, obs)
      int64
      2 1 3 1 0 2 3 1 ... 0 1 1 2 3 3 0 2

      array([[[2, 1, 3, ..., 1, 3, 3],
              [0, 0, 0, ..., 0, 0, 0],
              [2, 0, 2, ..., 3, 2, 2],
              ...,
              [3, 3, 2, ..., 3, 3, 3],
              [1, 0, 1, ..., 3, 1, 1],
              [1, 3, 3, ..., 3, 0, 2]]])

  • Indexes: (3)
    • chain
      PandasIndex

      PandasIndex(Int64Index([0], dtype='int64', name='chain'))

    • draw
      PandasIndex

      PandasIndex(Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
                  ...
                  490, 491, 492, 493, 494, 495, 496, 497, 498, 499],
                 dtype='int64', name='draw', length=500))

    • obs
      PandasIndex

      PandasIndex(Int64Index([   0,    1,    2,    3,    4,    5,    6,    7,    8,    9,
                  ...
                  3146, 3147, 3148, 3149, 3150, 3151, 3152, 3153, 3154, 3155],
                 dtype='int64', name='obs', length=3156))

  • Attributes: (4)

    created_at :

    2023-07-13T17:31:04.377859

    arviz_version :

    0.15.1

    inference_library :

    pymc

    inference_library_version :

    5.3.0

  
  
• observed_data

  <xarray.Dataset>
  Dimensions:  (obs: 3156)
  Coordinates:
    * obs      (obs) int64 0 1 2 3 4 5 6 7 ... 3149 3150 3151 3152 3153 3154 3155
  Data variables:
      y_cat    (obs) int64 1 3 1 1 1 3 1 1 1 1 1 1 1 ... 2 2 3 2 2 2 2 2 2 2 2 2 2
  Attributes:
      created_at:                 2023-07-13T17:31:04.378101
      arviz_version:              0.15.1
      inference_library:          pymc
      inference_library_version:  5.3.0

  xarray.Dataset

  • Dimensions:
    • obs: 3156
  • Coordinates: (1)
    • obs
      (obs)
      int64
      0 1 2 3 4 ... 3152 3153 3154 3155

      array([   0,    1,    2, ..., 3153, 3154, 3155])

  • Data variables: (1)
    • y_cat
      (obs)
      int64
      1 3 1 1 1 3 1 1 ... 2 2 2 2 2 2 2 2

      array([1, 3, 1, ..., 2, 2, 2])

  • Indexes: (1)
    • obs
      PandasIndex

      PandasIndex(Int64Index([   0,    1,    2,    3,    4,    5,    6,    7,    8,    9,
                  ...
                  3146, 3147, 3148, 3149, 3150, 3151, 3152, 3153, 3154, 3155],
                 dtype='int64', name='obs', length=3156))

  • Attributes: (4)

    created_at :

    2023-07-13T17:31:04.378101

    arviz_version :

    0.15.1

    inference_library :

    pymc

    inference_library_version :

    5.3.0

  
  

beta_individual = idata_m4["posterior"]["beta_individual"]
predicted = beta_individual.mean(("chain", "draw"))
predicted = predicted.sortby(predicted.sel(alts_intercepts="nabisco"))
ci_lb = beta_individual.quantile(0.025, ("chain", "draw")).sortby(
    predicted.sel(alts_intercepts="nabisco")
)
ci_ub = beta_individual.quantile(0.975, ("chain", "draw")).sortby(
    predicted.sel(alts_intercepts="nabisco")
)

fig = plt.figure(figsize=(10, 9))
gs = fig.add_gridspec(
    2,
    3,
    width_ratios=(4, 4, 4),
    height_ratios=(1, 7),
    left=0.1,
    right=0.9,
    bottom=0.1,
    top=0.9,
    wspace=0.05,
    hspace=0.05,
)
# Create the Axes.
ax = fig.add_subplot(gs[1, 0])
ax.set_yticklabels([])
ax_histx = fig.add_subplot(gs[0, 0], sharex=ax)
ax_histx.set_title("Expected Modifications \n to Nabisco Baseline", fontsize=10)
ax_histx.hist(predicted.sel(alts_intercepts="nabisco"), bins=30, ec="black", color="red")
ax_histx.set_yticklabels([])
ax_histx.tick_params(labelsize=8)
ax.set_ylabel("Individuals", fontsize=10)
ax.tick_params(labelsize=8)
ax.hlines(
    range(len(predicted)),
    ci_lb.sel(alts_intercepts="nabisco"),
    ci_ub.sel(alts_intercepts="nabisco"),
    color="black",
    alpha=0.3,
)
ax.scatter(predicted.sel(alts_intercepts="nabisco"), range(len(predicted)), color="red", ec="white")
ax.fill_betweenx(range(139), -0.03, 0.03, alpha=0.2, color="red")

ax1 = fig.add_subplot(gs[1, 1])
ax1.set_yticklabels([])
ax_histx = fig.add_subplot(gs[0, 1], sharex=ax1)
ax_histx.set_title("Expected Modifications \n to Keebler Baseline", fontsize=10)
ax_histx.set_yticklabels([])
ax_histx.tick_params(labelsize=8)
ax_histx.hist(predicted.sel(alts_intercepts="keebler"), bins=30, ec="black", color="red")
ax1.hlines(
    range(len(predicted)),
    ci_lb.sel(alts_intercepts="keebler"),
    ci_ub.sel(alts_intercepts="keebler"),
    color="black",
    alpha=0.3,
)
ax1.scatter(
    predicted.sel(alts_intercepts="keebler"), range(len(predicted)), color="red", ec="white"
)
ax1.set_xlabel("Individual Modifications to the Product Intercept", fontsize=10)
ax1.fill_betweenx(range(139), -0.03, 0.03, alpha=0.2, color="red", label="Negligible \n Region")
ax1.tick_params(labelsize=8)
ax1.legend(fontsize=10)

ax2 = fig.add_subplot(gs[1, 2])
ax2.set_yticklabels([])
ax_histx = fig.add_subplot(gs[0, 2], sharex=ax2)
ax_histx.set_title("Expected Modifications \n to Sunshine Baseline", fontsize=10)
ax_histx.set_yticklabels([])
ax_histx.hist(predicted.sel(alts_intercepts="sunshine"), bins=30, ec="black", color="red")
ax2.hlines(
    range(len(predicted)),
    ci_lb.sel(alts_intercepts="sunshine"),
    ci_ub.sel(alts_intercepts="sunshine"),
    color="black",
    alpha=0.3,
)
ax2.fill_betweenx(range(139), -0.03, 0.03, alpha=0.2, color="red")
ax2.scatter(
    predicted.sel(alts_intercepts="sunshine"), range(len(predicted)), color="red", ec="white"
)
ax2.tick_params(labelsize=8)
ax_histx.tick_params(labelsize=8)
plt.suptitle("Individual Differences by Product", fontsize=20);

This type of plot is often useful for identifying loyal customers. Similarly it can be used to identify cohorts of customers that ought to be better incentivised if we hope them to switch to our product.

Conclusion

We can see here the flexibility and richly parameterised possibilities for modelling individual choice of discrete options. These techniques are useful in a wide variety of domains from microeconomics, to marketing and product development. The notions of utility, probability and their interaction lie at the heart of Savage's Representation theorem and justification(s) for Bayesian approaches to statistical inference. So discrete modelling is a natural fit for the Bayesian, but Bayesian statistics is also a natural fit for discrete choice modelling. The traditional estimation techniques are often brittle and very dependent on starting values of the MLE process. The Bayesian setting trades this brittleness for a framework which allows us to incorporate our beliefs about what drives human utility calculations. We've only scratched the surface in this example notebook, but encourage you to further explore the technique.

Authors

• Authored by Nathaniel Forde in June 2023

References

:::{bibliography} :filter: docname in docnames :::

Watermark

%load_ext watermark
%watermark -n -u -v -iv -w -p pytensor

Last updated: Thu Jul 13 2023

Python implementation: CPython
Python version       : 3.9.16
IPython version      : 8.11.0

pytensor: 2.11.1

numpy     : 1.23.5
matplotlib: 3.7.1
pytensor  : 2.11.1
pandas    : 1.5.3
arviz     : 0.15.1
pymc      : 5.3.0

Watermark: 2.3.1

:::{include} ../page_footer.md :::