inst/examples/swissmetro/plot_b26_triangular_panel_mixture.R

In rbiogeme: Interface to 'Biogeme'

#!/usr/bin/env Rscript

# b26. Triangular mixture with panel data
#
# This example mirrors plot_b26_triangular_panel_mixture.py. The same
# individual-level triangular random coefficients are shared across each
# person's choice trajectory, and native PanelLikelihoodTrajectory aggregates
# the conditional probabilities before Monte Carlo integration.

library(rbiogeme)

# prepare_swissmetro_example() is defined in example_utils.R. It parses the
# command line, validates the data/Python paths, configures the bridge, reads
# the data, and creates a fresh output directory. It does not define the model.
script_path <- commandArgs(trailingOnly = FALSE)
script_path <- sub("^--file=", "", script_path[startsWith(script_path, "--file=")][[1L]])
source(file.path(dirname(normalizePath(script_path)), "example_utils.R"))

build_b26_triangular_panel_mixture_model <- function(database) {
  # All parameter names and starting values match the native panel example.
  b_cost <- biogeme_beta("b_cost", start = 0)
  b_time <- biogeme_beta("b_time", start = 0)
  b_time_s <- biogeme_beta("b_time_s", start = 1)
  asc_car <- biogeme_beta("asc_car", start = 0)
  asc_car_s <- biogeme_beta("asc_car_s", start = 1)
  asc_train <- biogeme_beta("asc_train", start = 0)
  asc_train_s <- biogeme_beta("asc_train_s", start = 1)
  asc_sm <- biogeme_beta("asc_sm", start = 0, fixed = TRUE)
  asc_sm_s <- biogeme_beta("asc_sm_s", start = 1)

  # Each draw type uses the same bridge-owned native triangular generator.
  b_time_draws <- biogeme_draws(
    "b_time_rnd", "TRIANGULAR", number_of_draws = 5000, seed = 1223,
    generator = "TRIANGULAR"
  )
  asc_car_draws <- biogeme_draws(
    "asc_car_rnd", "TRIANGULAR", number_of_draws = 5000, seed = 1223,
    generator = "TRIANGULAR"
  )
  asc_train_draws <- biogeme_draws(
    "asc_train_rnd", "TRIANGULAR", number_of_draws = 5000, seed = 1223,
    generator = "TRIANGULAR"
  )
  asc_sm_draws <- biogeme_draws(
    "asc_sm_rnd", "TRIANGULAR", number_of_draws = 5000, seed = 1223,
    generator = "TRIANGULAR"
  )
  b_time_rnd <- b_time + b_time_s * b_time_draws
  asc_car_rnd <- asc_car + asc_car_s * asc_car_draws
  asc_train_rnd <- asc_train + asc_train_s * asc_train_draws
  asc_sm_rnd <- asc_sm + asc_sm_s * asc_sm_draws

  v_train <- asc_train_rnd + b_time_rnd * variable("TRAIN_TT_SCALED") +
    b_cost * variable("TRAIN_COST_SCALED")
  v_swissmetro <- asc_sm_rnd + b_time_rnd * variable("SM_TT_SCALED") +
    b_cost * variable("SM_COST_SCALED")
  v_car <- asc_car_rnd + b_time_rnd * variable("CAR_TT_SCALED") +
    b_cost * variable("CAR_CO_SCALED")
  conditional_probability <- logit_probability(
    utilities = list(`1` = v_train, `2` = v_swissmetro, `3` = v_car),
    availability = list(
      `1` = variable("TRAIN_AV_SP"),
      `2` = variable("SM_AV"),
      `3` = variable("CAR_AV_SP")
    ),
    alternative = variable("CHOICE")
  )
  trajectory <- panel_likelihood_trajectory(conditional_probability)
  log_probability <- log(monte_carlo(trajectory))

  biogeme_model(
    database = database,
    formula = log_probability,
    draws = list(b_time_draws, asc_car_draws, asc_train_draws, asc_sm_draws),
    control = biogeme_control(
    output_directory = prepared$output,
      model_name = "b26_triangular_panel_mixture",
      number_of_draws = 5000,
      seed = 1223,
      calculating_second_derivatives = "never",
      generate_html = FALSE,
      generate_yaml = FALSE,
      save_iterations = FALSE
    )
  )
}

prepared <- prepare_swissmetro_example(
  commandArgs(trailingOnly = TRUE),
  default_model = "b26_triangular_panel_mixture"
)

database <- swissmetro_data(prepared$data, panel = TRUE)
model <- build_b26_triangular_panel_mixture_model(database)
fit <- estimate(
  model,
  model_name = "b26_triangular_panel_mixture",
  control = model$control
)

print(summary(fit))
print(coef(fit))

invisible(fit)