R: Distributional Likelihood for Hurdle Model Count Data...

dist_ll {hurdlr}

R Documentation

Distributional Likelihood for Hurdle Model Count Data Regression

dist_ll is the data likelihood fuction for hurdle model regression using hurdle.

dist_ll(y, hurd = Inf, lam = NULL, size = 1, mu = NULL, xi = NULL,
  sigma = NULL, dist = c("poisson", "nb", "lognormal", "gpd"), g.x = F,
  log = T)

`y`	numeric response vector.
`hurd`	numeric threshold for 'extreme' observations of two-hurdle models. `Inf` for one-hurdle models.
`lam`	current value for the poisson likelihood lambda parameter.
`size`	size parameter for negative binomial likelihood distributions.
`mu`	current value for the negative binomial or log normal likelihood mu parameter.
`xi`	current value for the generalized pareto likelihood xi parameter.
`sigma`	current value for the generalized pareto likelihood sigma parameter.
`dist`	character specification of response distribution.
`g.x`	logical operator. `TRUE` if operating within the third component of the likelihood function (the likelihood of 'extreme' observations).
`log`	logical operator. if `TRUE`, probabilities p are given as log(p).

Currently, Poisson, Negative Binomial, log-Normal, and Generalized Pareto distributions are available.

The log-likelihood of the zero-inflated Poisson fit for the current iteration of the MCMC algorithm.

Taylor Trippe <ttrippe@luc.edu>
Earvin Balderama <ebalderama@luc.edu>