Step 1: Building reinforcement learning model

Usage

run_m(
  data,
  colnames = list(),
  behrule = list(),
  funcs = list(),
  params = list(),
  priors = list(),
  settings = list(),
  engine = "Cpp",
  ...
)

Arguments

data: A data frame in which each row represents a single trial, see data
colnames: Column names in the data frame, see colnames
behrule: The agent’s implicitly formed internal rule, see behrule
funcs: The functions forming the reinforcement learning model, see funcs
params: Parameters used by the model’s internal functions, see params
priors: Prior probability density function of the free parameters, see priors
settings: Other model settings, see settings
engine: Specifies whether the core Markov Decision Process (MDP) update loop is executed in C++ or in R.
...: Additional arguments passed to internal functions.

Value

An S4 object of class multiRL.model.

input: An S4 object of class multiRL.input, containing the raw data, column specifications, parameters and ...
behrule: An S4 object of class multiRL.behrule, defining the latent learning rules.
result: An S4 object of class multiRL.result, storing trial-level outputs of the Markov Decision Process.
sumstat: An S4 object of class multiRL.sumstat, providing summary statistics across different estimation methods.
extra: A List containing additional user-defined information.

Examples

multiRL.model <- multiRL::run_m(
  data = multiRL::TAB[multiRL::TAB[, "Subject"] == 1, ],
  behrule = list(
    cue = c("A", "B", "C", "D"),
    rsp = c("A", "B", "C", "D")
  ),
  colnames = list(
    subid = "Subject", block = "Block", trial = "Trial",
    object = c("L_choice", "R_choice"), 
    reward = c("L_reward", "R_reward"),
    action = "Sub_Choose",
    exinfo = c("Frame", "NetWorth", "RT")
  ),
  params = list(
    free = list(
      alpha = 0.5,
      beta = 0.5
    ),
    fixed = list(
      gamma = 1, 
      delta = 0.1, 
      epsilon = NA_real_,
      zeta = 0
    ),
    constant = list(
      seed = 123,
      Q0 = NA_real_, 
      reset = NA_real_,
      lapse = 0.01,
      threshold = 1,
      bonus = 0,
      weight = 1,
      capacity = 0,
      sticky = 0
    )
  ),
  priors = list(
    alpha = function(x) {stats::dbeta(x, shape1 = 2, shape2 = 2, log = TRUE)}, 
    beta = function(x) {stats::dexp(x, rate = 1, log = TRUE)}
  ),
  settings = list(
    name = "TD",
    mode = "fitting",
    estimate = "MLE",
    policy = "off",
    system = c("RL", "WM")
 ),
 engine = "R"
)
 
multiRL.summary <- multiRL::summary(multiRL.model)
#> Model Fit:
#>   Accuracy: 100%
#>   Log-Likelihood: -369.89
#>   Log-Prior Probability: -0.09
#>   Log-Posterior Probability: -369.99
#>   AIC: 743.79
#>   BIC: 751.56