The {choicedata} package simplifies working with choice
data in R.
The package breaks down the process of modeling choice data into a series of steps. Each step is represented by an object that contains the necessary information for the subsequent step.
choice_formula:
the choice model formula.
choice_alternatives:
the set of choice alternatives.
choice_effects:
the choice effects, defined by choice_alternatives and
choice_formula.
choice_parameters:
the model parameters, determined by choice_effects and
estimated via choice_likelihood.
choice_identifiers:
the identifiers for deciders and choice occasions.
choice_preferences:
the deciders’ choice preferences, identified by
choice_identifiers.
choice_responses:
the choice responses, influenced by
choice_preferences.
choice_covariates:
the choice covariates.
choice_data:
the choice data, built by choice_covariates and
choice_responses.
choice_probabilities:
the choice probabilities, computed from choice_data and
choice_parameters.
choice_likelihood:
the likelihood of the choice model, formed by
choice_probabilities.
The objects are designed to be modular and can be combined in various ways to create a range of modeling workflows.
The travel_mode_choice data set contains the revealed
preferences of 210 travelers choosing between plane, train, bus, and
car. We can transform the data from long to wide format, or construct
model design matrices:
library("choicedata")
travel_mode_choice
#> # A tibble: 840 × 8
#> individual mode choice wait cost travel income size
#> <int> <chr> <int> <int> <int> <int> <int> <int>
#> 1 1 plane 0 69 59 100 35 1
#> 2 1 train 0 34 31 372 35 1
#> 3 1 bus 0 35 25 417 35 1
#> 4 1 car 1 0 10 180 35 1
#> 5 2 plane 0 64 58 68 30 2
#> 6 2 train 0 44 31 354 30 2
#> 7 2 bus 0 53 25 399 30 2
#> 8 2 car 1 0 11 255 30 2
#> 9 3 plane 0 69 115 125 40 1
#> 10 3 train 0 34 98 892 40 1
#> # ℹ 830 more rows
long_to_wide(
data_frame = travel_mode_choice,
column_alternative = "mode",
column_decider = "individual"
)
#> # A tibble: 210 × 16
#> individual income size wait_plane wait_train wait_bus wait_car cost_plane
#> <int> <int> <int> <int> <int> <int> <int> <int>
#> 1 1 35 1 69 34 35 0 59
#> 2 2 30 2 64 44 53 0 58
#> 3 3 40 1 69 34 35 0 115
#> 4 4 70 3 64 44 53 0 49
#> 5 5 45 2 64 44 53 0 60
#> 6 6 20 1 69 40 35 0 59
#> 7 7 45 1 45 34 35 0 148
#> 8 8 12 1 69 34 35 0 121
#> 9 9 40 1 69 34 35 0 59
#> 10 10 70 2 69 34 35 0 58
#> # ℹ 200 more rows
#> # ℹ 8 more variables: cost_train <int>, cost_bus <int>, cost_car <int>,
#> # travel_plane <int>, travel_train <int>, travel_bus <int>, travel_car <int>,
#> # choice <chr>
mode_effects <- choice_effects(
choice_formula = choice_formula(
formula = choice ~ cost | income + size | travel + wait,
error_term = "probit"
),
choice_alternatives = choice_alternatives(
J = 4,
alternatives = unique(travel_mode_choice$mode)
)
)
mode_data <- choice_data(
data_frame = travel_mode_choice,
format = "long",
column_choice = "choice",
column_decider = "individual",
column_alternative = "mode",
column_ac_covariates = c("income", "size"),
column_as_covariates = c("wait", "cost", "travel")
)
design_matrices <- design_matrices(mode_data, mode_effects)
design_matrices[[1]]
#> cost income_car income_plane income_train size_car size_plane size_train
#> bus 25 0 0 0 0 0 0
#> car 10 35 0 0 1 0 0
#> plane 59 0 35 0 0 1 0
#> train 31 0 0 35 0 0 1
#> ASC_car ASC_plane ASC_train travel_bus travel_car travel_plane
#> bus 0 0 0 417 0 0
#> car 1 0 0 0 180 0
#> plane 0 1 0 0 0 100
#> train 0 0 1 0 0 0
#> travel_train wait_bus wait_car wait_plane wait_train
#> bus 0 35 0 0 0
#> car 0 0 0 0 0
#> plane 0 0 0 69 0
#> train 372 0 0 0 34generate_choice_data() makes it straightforward to
simulate choice data. The example below simulates 200 ranking tasks with
three alternatives and recovers the data-generating parameters via
numerical optimization of the likelihood:
library("choicedata")
set.seed(1)
sim_effects <- choice_effects(
choice_formula = choice_formula(
formula = choice ~ A | B | C,
error_term = "logit"
),
choice_alternatives = choice_alternatives(
J = 3,
alternatives = c("A", "B", "C")
)
)
sim_parameters <- generate_choice_parameters(sim_effects)
(sim_data <- generate_choice_data(
choice_effects = sim_effects,
choice_identifiers = generate_choice_identifiers(N = 200),
choice_parameters = sim_parameters,
choice_type = "ranked"
))
#> # A tibble: 200 × 13
#> deciderID occasionID choice B choice_A choice_B choice_C A_A
#> * <chr> <chr> <chr> <dbl> <int> <int> <int> <dbl>
#> 1 1 1 B 1.12 2 1 3 0.576
#> 2 2 1 B 0.782 3 1 2 -0.0449
#> 3 3 1 B -0.478 2 1 3 0.0746
#> 4 4 1 B -0.415 3 1 2 0.418
#> 5 5 1 B 0.697 2 1 3 -0.394
#> 6 6 1 B 0.881 2 1 3 0.557
#> 7 7 1 B -0.367 3 1 2 0.398
#> 8 8 1 B 0.0280 3 1 2 -1.04
#> 9 9 1 C 0.476 3 2 1 -0.743
#> 10 10 1 B 0.00111 2 1 3 -0.710
#> # ℹ 190 more rows
#> # ℹ 5 more variables: A_B <dbl>, A_C <dbl>, C_A <dbl>, C_B <dbl>, C_C <dbl>
sim_likelihood <- choice_likelihood(
choice_data = sim_data,
choice_effects = sim_effects
)
objective <- sim_likelihood$objective
true_vector <- switch_parameter_space(
choice_parameters = sim_parameters,
choice_effects = sim_effects
)
fit <- stats::optim(
par = stats::rnorm(length(true_vector)),
fn = function(par) {
objective(choice_parameters = par, logarithm = TRUE, negative = TRUE)
}
)
estimated_parameters <- switch_parameter_space(
choice_parameters = fit$par,
choice_effects = sim_effects
)
data.frame(dgp = true_vector, estimated = fit$par)
#> dgp estimated
#> beta_1 -1.9810209 -2.04179163
#> beta_2 0.5807312 0.04789289
#> beta_3 -2.6424897 -3.05406623
#> beta_4 5.0447208 4.82873952
#> beta_5 1.0419951 0.99061141
#> beta_6 -2.5945488 -2.75745549
#> beta_7 1.5413860 1.21554512
#> beta_8 2.3347877 1.96729481You can install the released package version from CRAN with:
install.packages("choicedata"){Rprobit} (Bauer et
al. 2023) provides maximum approximated composite marginal
likelihood estimation for efficient probit choice modeling.
{RprobitB} (Oelschläger
and Bauer 2025) provides Bayesian tools for estimating probit
models.
You have a question, found a bug, request a feature, want to give feedback, or like to contribute? Please file an issue on GitHub.