Shifted Inverse Weibull Model (Fréchet Distribution)

Provides the necessary functions to use a shifted Inverse Weibull distribution as a custom family in brms. This version is reparameterized to use tau (proportion of non-decision time relative to minimum RT) and minrt (minimum possible RT), where the non-decision time ndt = tau * minrt. The distribution is parameterized by the shape (mu), scale (sigma), as well as tau, and minrt (which is typically fixed).

Usage

rt_invweibull(
  link_mu = "softplus",
  link_sigma = "softplus",
  link_tau = "logit",
  link_minrt = "identity"
)

rt_invweibull_lpdf_expose()

rt_invweibull_stanvars()

log_lik_rt_invweibull(i, prep)

posterior_predict_rt_invweibull(i, prep, ...)

posterior_epred_rt_invweibull(prep)

Arguments

link_mu, link_sigma, link_tau, link_minrt

Link function for the parameters.

i, prep

For brms' functions to run: index of the observation and a brms preparation object.

...

Additional arguments.