inst/examples/swissmetro/plot_b06b_unif_mixture_MHLS.R

In rbiogeme: Interface to 'Biogeme'

#!/usr/bin/env Rscript

# b06b. Uniform mixture with normal MLHS draws
#
# This example estimates a random time coefficient with native modified Latin
# hypercube sampling. The draw type is NORMAL_MLHS, matching the Python
# example; the likelihood and all numerical work remain in Biogeme.

library(rbiogeme)

# The shared helper contains command-line parsing and data preparation. The
# complete model specification remains in this script.
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_b06b_unif_mixture_mlhs_model <- function(
    database,
    number_of_draws = 10000L,
    seed = 1223L
) {
  # These parameter names and starting values match native b06b. The
  # Swissmetro ASC is fixed at zero to identify the utility scale.
  asc_car <- biogeme_beta("asc_car", start = 0)
  asc_train <- biogeme_beta("asc_train", start = 0)
  asc_sm <- biogeme_beta("asc_sm", start = 0, fixed = TRUE)
  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)

  # NORMAL_MLHS is native modified Latin hypercube sampling from a normal
  # distribution. draw() is a symbolic native node, not an R callback.
  b_time_rnd <- b_time + b_time_s * draw("b_time_rnd", "NORMAL_MLHS")

  # Utilities and availability are symbolic R expressions. They are compiled
  # once into native Biogeme before numerical evaluation starts.
  utilities <- list(
    `1` = asc_train + b_time_rnd * variable("TRAIN_TT_SCALED") +
      b_cost * variable("TRAIN_COST_SCALED"),
    `2` = asc_sm + b_time_rnd * variable("SM_TT_SCALED") +
      b_cost * variable("SM_COST_SCALED"),
    `3` = asc_car + b_time_rnd * variable("CAR_TT_SCALED") +
      b_cost * variable("CAR_CO_SCALED")
  )
  availability <- list(
    `1` = variable("TRAIN_AV_SP"),
    `2` = variable("SM_AV"),
    `3` = variable("CAR_AV_SP")
  )

  # The observed-choice selector compiles to native models.logit(...,
  # i=CHOICE). Monte Carlo integration is also a native expression node.
  conditional_probability <- logit_probability(
    utilities = utilities,
    availability = availability,
    alternative = variable("CHOICE")
  )
  log_likelihood <- log(monte_carlo(conditional_probability))
  draws <- biogeme_draws(
    name = "b_time_rnd",
    draw_type = "NORMAL_MLHS",
    number_of_draws = number_of_draws,
    seed = seed
  )

  biogeme_model(
    database = database,
    formula = log_likelihood,
    draws = draws
  )
}

# 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. The --data, --python,
# --output, --draws, and --seed options work from any current working
# directory.
prepared <- prepare_swissmetro_example(
  commandArgs(trailingOnly = TRUE),
  default_model = "b06b_unif_mixture_MHLS"
)

number_of_draws <- if (!is.null(prepared$options$draws) && nzchar(prepared$options$draws)) {
  example_integer(prepared$options$draws, "draws")
} else {
  10000L
}
seed <- if (!is.null(prepared$options$seed) && nzchar(prepared$options$seed)) {
  example_integer(prepared$options$seed, "seed")
} else {
  1223L
}

# estimate() always performs fresh native estimation. Remove only exact b06b
# artifacts so an old YAML or iteration file cannot silently be reused.
stale_files <- c(
  "b06b_unif_mixture_MHLS.yaml",
  "__b06b_unif_mixture_MHLS.iter",
  "b06b_unif_mixture_MHLS.html"
)
stale_files <- file.path(prepared$output, stale_files)
stale_files <- stale_files[file.exists(stale_files)]
if (length(stale_files) > 0L) unlink(stale_files, force = TRUE)

database <- swissmetro_data(prepared$data)
model <- build_b06b_unif_mixture_mlhs_model(
  database,
  number_of_draws = number_of_draws,
  seed = seed
)
control <- biogeme_control(
    output_directory = prepared$output,
  model_name = "b06b_unif_mixture_MHLS",
  number_of_draws = number_of_draws,
  seed = seed,
  analytical_hessian_mode = "automatic",
  generate_html = TRUE,
  generate_yaml = FALSE,
  save_iterations = FALSE
)

cat(sprintf("Number of draws: %s\n", format(number_of_draws, big.mark = "_")))

# The complete MLHS-aware expression graph is compiled once. Native Python
# Biogeme performs the draw generation, Monte Carlo integration, and estimate.
fit <- estimate(
  model,
  model_name = "b06b_unif_mixture_MHLS",
  control = control
)

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