bayesPF: Fits a psychometric function

In adknudson/BayesPsychometric: Fits Psychometric Functions Using Bayesian Methods

Description

Fits a psychometric function

Usage

bayesPF(formula, data, link, adaptive_pooling = FALSE, chains = 1,
  iter = 2000, warmup = 1000, thin = 1, cores = 1, ...,
  sample = TRUE)

Arguments

formula	A formula that one would pass to 'glm' or similar
data	A data.frame object
link	A link function (either "logit" or "probit")
adaptive_pooling	Logical (FALSE by default) Specifies whether adaptive pooling should be used when fitting the model. Requires that both an intecept and at least one factor variable is included in the formula.
...	Other parameters to be passed to 'rstan::stan()'.
sample	Logical. Whether to sample from the model or just build the model without sampling

Examples

# Load the data
data_binomial  <- Sample_Data_Binomial
data_bernoulli <- Sample_Data_Bernoulli

# Fit a full model using binomial response data. By default, it will only
# return the posterior samples for the coefficients
fit1.1 <- bayesPF(y|k ~ x1 + age + gender, data_binomial, "logit",
                  chains = 2, cores = 2, iter = 8000, warmup = 2000)

str(fit1.1)

# Fit a full model using binary response data. To get the fitted Stan
# model, specify `return_stan_fit = TRUE`. To get diagnostic information
# such as the Stan code and transformed data, specify `return_f2stan = TRUE`
fit1.2 <- bayesPF(y ~ x1 + age + gender, data_bernoulli, "probit",
                  return_f2stan = TRUE, return_stan_fit = TRUE)
str(fit1.2)
extract(fit1.2$fit)

adknudson/BayesPsychometric documentation built on Nov. 22, 2019, 1:59 p.m.