inst/examples/swissmetro/plot_b09_nested.R

In rbiogeme: Interface to 'Biogeme'

#!/usr/bin/env Rscript

# b09. Nested logit model
#
# This example mirrors plot_b09_nested.py. Train and Car share a non-trivial
# nest called "existing"; Swissmetro remains a trivial one-alternative nest.

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. The --data, --python, and
# --output options work from any current working directory.
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_b09_nested_model <- function(database) {
  # These parameter names, starting values, bounds, and fixed ASC match the
  # native Python example exactly.
  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_time <- biogeme_beta("b_time", start = 0)
  b_cost <- biogeme_beta("b_cost", start = 0)
  nest_parameter <- biogeme_beta(
    "nest_parameter",
    start = 1,
    lower = 1,
    upper = 3
  )

  utilities <- list(
    `1` = asc_train + b_time * variable("TRAIN_TT_SCALED") +
      b_cost * variable("TRAIN_COST_SCALED"),
    `2` = asc_sm + b_time * variable("SM_TT_SCALED") +
      b_cost * variable("SM_COST_SCALED"),
    `3` = asc_car + b_time * 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")
  )

  # nested_nest() mirrors OneNestForNestedLogit. Alternatives not listed in a
  # non-trivial nest become native trivial nests automatically.
  existing <- nested_nest(
    nest_parameter = nest_parameter,
    alternatives = c(1L, 3L),
    name = "existing"
  )
  nests <- nested_nests(
    choice_set = c(1L, 2L, 3L),
    nests = list(existing)
  )

  # nested_logit_model() compiles native models.lognested and preserves the
  # nest parameter and alternative membership in the complete model object.
  nested_logit_model(
    database = database,
    choice = "CHOICE",
    utilities = utilities,
    availability = availability,
    nests = nests,
    control = biogeme_control(
    output_directory = prepared$output,
      model_name = "b09_nested",
      optimization_algorithm = "simple_bounds_BFGS",
      generate_html = TRUE,
      generate_yaml = FALSE,
      save_iterations = FALSE
    )
  )
}

prepared <- prepare_swissmetro_example(
  commandArgs(trailingOnly = TRUE),
  default_model = "b09_nested"
)

# Always estimate from the expression tree. Remove only exact b09 artifacts so
# an old YAML or iteration file cannot silently be recycled.
stale_files <- c(
  "b09_nested.yaml",
  "__b09_nested.iter",
  "b09_nested.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_b09_nested_model(database)

# estimate() delegates the nested likelihood, derivatives, bounded
# optimization, null-likelihood reporting, and result reporting to Biogeme.
fit <- estimate(
  model,
  model_name = "b09_nested",
  control = model$control
)

# The native example reports the correlation of the alternative error terms
# after estimation. This call uses native NestsForNestedLogit.correlation().
correlation <- nested_logit_correlation(
  model,
  beta_values = coef(fit),
  alternatives_names = c(`1` = "Train", `2` = "Swissmetro", `3` = "Car")
)

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