estimate_twochoiceRL: Estimate behavior in a two-choice decision task

In psuthaharan/twochoiceRL: Modeling Behavior for Two-Choice Decisions

Description

This function performs parameter estimation to recover behavior in a two-choice decision task

Usage

estimate_twochoiceRL(
  seed_value = 528,
  data = NULL,
  method = "mle",
  nRes = 5,
  tr_rad = c(1, 5),
  prior_mean = c(NULL, NULL),
  prior_sd = c(NULL, NULL),
  plot = FALSE,
  progress_bar = TRUE
)

Arguments

seed_value	user-specified random value that ensures replication of results. Defaults to 528.
data	simulated two-choice task data
method	parameter estimation technique used; either maximum-likelihood estimation (mle), maximum a posteriori (map) or expectation-maximization with laplace approximation (eml). Defaults to mle.
nRes	number of restarts. Defaults to 5.
tr_rad	starting and maximum allowed trust region radius. Defaults to 1 and 5.
prior_mean	initial prior means of parameters (x1,x2). Set value if method = "map" or "eml".
prior_sd	initial prior standard deviations of parameters (x1,x2). Set value if method = "map" or "eml".
plot	visualize estimation performance between true parameters vs estimated parameters. Defaults to FALSE.
progress_bar	track completion time of estimation procedure. Defaults to TRUE.

Value

A list containing a dataframe of the true and estimated parameter values, the measures of the bias, rmse and pearson correlation of parameters x1 and x2, the posterior hyperparameters (if method = "eml") per iteration, plot (if plot = TRUE) of the relationship between the true and estimated parameter values, and a plot (if method = eml and plot = TRUE) of the posterior hyperparameters per iteration.

Examples

# Save simulated task data to a variable, say, data_sim
data_sim <- simulate_twochoiceRL(trials_unique = TRUE)

# Recover behavioral parameters using maximum-likelihood estimation (MLE)
est_mle <- estimate_twochoiceRL(data = data_sim, method = "mle", plot=FALSE)

# View the true and MLE-estimated parameter values
View(est_mle[[1]]$value)

# If plot=TRUE, view correlation plot between the true and MLE-estimated parameter for x1
View(est_mle[[2]])

psuthaharan/twochoiceRL documentation built on Dec. 22, 2021, 9:56 a.m.