zero_nb: Zero-Inflated Negative Binomial Regression Model
In hurdlr: Zero-Inflated and Hurdle Modelling Using Bayesian Inference

Description Usage Arguments Details Value Author(s)

zero_nb is used to fit zero-inflated negative binomial regression models to count data via Bayesian inference.

1
2
3

zero_nb(y, x, size, a = 1, b = 1, mu.start = 1, beta.prior.mean = 0,
  beta.prior.sd = 1, iters = 1000, burn = 500, nthin = 1, plots = T,
  progress.bar = T)

`y`	numeric response vector.
`x`	numeric predictor matrix.
`size`	size parameter for negative binomial likelihood distributions.
`a`	shape parameter for gamma prior distributions.
`b`	rate parameter for gamma prior distributions.
`mu.start`	initial value for mu parameter.
`beta.prior.mean`	mu parameter for normal prior distributions.
`beta.prior.sd`	standard deviation for normal prior distributions.
`iters`	number of iterations for the Markov chain to run.
`burn`	numeric burn-in length.
`nthin`	numeric thinning rate.
`plots`	logical operator. `TRUE` to output plots.
`progress.bar`	logical operator. `TRUE` to print progress bar.

Fits a zero-inflated negative binomial (ZINB) model.

zero_nb returns a list which includes the items

mu: numeric vector; posterior distribution of mu parameter
beta: numeric matrix; posterior distributions of regression coefficients
p: numeric vector; posterior distribution of parameter 'p', the probability of a given zero observation belonging to the model's zero component
ll: numeric vector; posterior log-likelihood

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

hurdlr documentation built on May 2, 2019, 3:19 p.m.