17a. Mixture with lognormal distributionΒΆ

Example of a mixture of logit models, using Monte-Carlo integration. The mixing distribution is distributed as a log normal.

Michel Bierlaire, EPFL Thu Jun 26 2025, 15:31:41

from IPython.core.display_functions import display

import biogeme.biogeme_logging as blog
from biogeme.biogeme import BIOGEME
from biogeme.expressions import Beta, Draws, MonteCarlo, exp, log
from biogeme.models import logit
from biogeme.results_processing import (
    EstimationResults,
    get_pandas_estimated_parameters,
)

See the data processing script: Data preparation for Swissmetro.

from swissmetro_data import (
    CAR_AV_SP,
    CAR_CO_SCALED,
    CAR_TT_SCALED,
    CHOICE,
    SM_AV,
    SM_COST_SCALED,
    SM_TT_SCALED,
    TRAIN_AV_SP,
    TRAIN_COST_SCALED,
    TRAIN_TT_SCALED,
    database,
)

logger = blog.get_screen_logger(level=blog.INFO)
logger.info('Example b17lognormal_mixture.py')
Example b17lognormal_mixture.py

Parameters to be estimated.

asc_car = Beta('asc_car', 0, None, None, 0)
asc_train = Beta('asc_train', 0, None, None, 0)
asc_sm = Beta('asc_sm', 0, None, None, 1)
b_cost = Beta('b_cost', 0, None, None, 0)

Define a random parameter, normally distributed, designed to be used for Monte-Carlo simulation.

b_time = Beta('b_time', 0, None, None, 0)

It is advised not to use 0 as starting value for the following parameter.

b_time_s = Beta('b_time_s', 1, -2, 2, 0)

Define a random parameter, log normally distributed, designed to be used for Monte-Carlo simulation.

b_time_rnd = -exp(b_time + b_time_s * Draws('b_time_rnd', 'NORMAL'))

Definition of the utility functions.

v_train = asc_train + b_time_rnd * TRAIN_TT_SCALED + b_cost * TRAIN_COST_SCALED
v_swissmetro = asc_sm + b_time_rnd * SM_TT_SCALED + b_cost * SM_COST_SCALED
v_car = asc_car + b_time_rnd * CAR_TT_SCALED + b_cost * CAR_CO_SCALED

Associate utility functions with the numbering of alternatives.

v = {1: v_train, 2: v_swissmetro, 3: v_car}

Associate the availability conditions with the alternatives.

av = {1: TRAIN_AV_SP, 2: SM_AV, 3: CAR_AV_SP}

Conditional to b_time_rnd, we have a logit model (called the kernel).

conditional_probability = logit(v, av, CHOICE)

We integrate over b_time_rnd using Monte-Carlo.

log_probability = log(MonteCarlo(conditional_probability))

As the objective is to illustrate the syntax, we calculate the Monte-Carlo approximation with a small number of draws.

the_biogeme = BIOGEME(database, log_probability, number_of_draws=10_000, seed=1223)
the_biogeme.model_name = 'b17a_lognormal_mixture'
Biogeme parameters read from biogeme.toml.

Estimate the parameters.

try:
    results = EstimationResults.from_yaml_file(
        filename=f'saved_results/{the_biogeme.model_name}.yaml'
    )
except FileNotFoundError:
    results = the_biogeme.estimate()
*** Initial values of the parameters are obtained from the file __b17a_lognormal_mixture.iter
Cannot read file __b17a_lognormal_mixture.iter. Statement is ignored.
Starting values for the algorithm: {}
As the model is rather complex, we cancel the calculation of second derivatives. If you want to control the parameters, change the algorithm from "automatic" to "simple_bounds" in the TOML file.
Optimization algorithm: hybrid Newton/BFGS with simple bounds [simple_bounds]
** Optimization: BFGS with trust region for simple bounds
Iter.       asc_train          b_time        b_time_s          b_cost         asc_car     Function    Relgrad   Radius      Rho
    0               0               0               1               0               0      5.7e+03      0.096      0.5 -0.00064    -
    1            -0.5             0.5             1.5            -0.5             0.5      5.4e+03      0.042      0.5     0.39    +
    2           -0.34            0.69             1.5              -1               0      5.3e+03       0.04      0.5     0.34    +
    3           -0.34            0.69             1.5              -1               0      5.3e+03       0.04     0.25    -0.63    -
    4           -0.39            0.44             1.3            -1.2            0.25      5.2e+03      0.017     0.25     0.49    +
    5           -0.39            0.61             1.3            -1.5           0.084      5.2e+03     0.0097     0.25     0.25    +
    6           -0.39            0.61             1.3            -1.5           0.084      5.2e+03     0.0097     0.12     -2.4    -
    7           -0.39            0.61             1.3            -1.5           0.084      5.2e+03     0.0097    0.062    0.017    -
    8           -0.32            0.55             1.2            -1.4            0.15      5.2e+03     0.0067    0.062     0.58    +
    9           -0.39            0.61             1.2            -1.4            0.17      5.2e+03     0.0065    0.062     0.15    +
   10           -0.33            0.57             1.3            -1.3            0.17      5.2e+03     0.0029    0.062     0.39    +
   11           -0.33            0.57             1.3            -1.3            0.17      5.2e+03     0.0029    0.031     -1.2    -
   12           -0.36            0.57             1.2            -1.4             0.2      5.2e+03     0.0036    0.031     0.23    +
   13           -0.34            0.58             1.2            -1.4            0.17      5.2e+03     0.0014    0.031     0.61    +
   14           -0.34            0.58             1.2            -1.4            0.17      5.2e+03     0.0014    0.016     -3.7    -
   15           -0.34            0.58             1.2            -1.4            0.17      5.2e+03     0.0014   0.0078    -0.65    -
   16           -0.35            0.57             1.2            -1.4            0.17      5.2e+03    0.00074   0.0078     0.49    +
   17           -0.35            0.58             1.2            -1.4            0.17      5.2e+03    0.00022   0.0078     0.65    +
   18           -0.35            0.58             1.2            -1.4            0.17      5.2e+03    0.00022   0.0039    -0.37    -
   19           -0.35            0.58             1.2            -1.4            0.18      5.2e+03    0.00023   0.0039     0.34    +
   20           -0.35            0.58             1.2            -1.4            0.18      5.2e+03    0.00023    0.002    -0.78    -
   21           -0.35            0.57             1.2            -1.4            0.17      5.2e+03     0.0002    0.002     0.13    +
   22           -0.35            0.58             1.2            -1.4            0.17      5.2e+03    0.00013    0.002     0.51    +
   23           -0.35            0.58             1.2            -1.4            0.17      5.2e+03    0.00013  0.00098    -0.59    -
   24           -0.35            0.58             1.2            -1.4            0.17      5.2e+03    0.00013  0.00049   -0.037    -
   25           -0.35            0.57             1.2            -1.4            0.17      5.2e+03    3.6e-05  0.00049     0.65    +
   26           -0.35            0.57             1.2            -1.4            0.17      5.2e+03    3.6e-05  0.00024    0.074    -
   27           -0.35            0.57             1.2            -1.4            0.17      5.2e+03    1.6e-05  0.00024     0.61    +
   28           -0.35            0.57             1.2            -1.4            0.17      5.2e+03      2e-05  0.00024     0.41    +
   29           -0.35            0.58             1.2            -1.4            0.17      5.2e+03    7.6e-06  0.00024     0.37    +
   30           -0.35            0.58             1.2            -1.4            0.17      5.2e+03    4.8e-06  0.00024     0.43    +
Optimization algorithm has converged.
Relative gradient: 4.8025028162905605e-06
Cause of termination: Relative gradient = 4.8e-06 <= 6.1e-06
Number of function evaluations: 70
Number of gradient evaluations: 39
Number of hessian evaluations: 0
Algorithm: BFGS with trust region for simple bound constraints
Number of iterations: 31
Proportion of Hessian calculation: 0/19 = 0.0%
Optimization time: 0:01:21.466994
Calculate second derivatives and BHHH
File b17a_lognormal_mixture.html has been generated.
File b17a_lognormal_mixture.yaml has been generated.
print(results.short_summary())
Results for model b17a_lognormal_mixture
Nbr of parameters:              5
Sample size:                    6768
Excluded data:                  3960
Final log likelihood:           -5231.272
Akaike Information Criterion:   10472.54
Bayesian Information Criterion: 10506.64
pandas_results = get_pandas_estimated_parameters(estimation_results=results)
display(pandas_results)
        Name     Value  Robust std err.  Robust t-stat.  Robust p-value
0  asc_train -0.346430         0.073266       -4.728357    2.263435e-06
1     b_time  0.575014         0.071294        8.065422    6.661338e-16
2   b_time_s  1.239151         0.128334        9.655646    0.000000e+00
3     b_cost -1.381117         0.097788      -14.123575    0.000000e+00
4    asc_car  0.173944         0.062414        2.786920    5.321162e-03

Total running time of the script: (3 minutes 0.119 seconds)

Gallery generated by Sphinx-Gallery