# countdistr: Count Data Distributions In tscount: Analysis of Count Time Series

## Description

Density, distribution function, quantile function, random generation, standard deviation and Anscombe residuals for some count data distributions. These auxiliary functions are used by several functions of the `tscount` package.

## Usage

 ```1 2 3 4 5 6 7``` ```ddistr(x, meanvalue, distr=c("poisson", "nbinom"), distrcoefs, ...) pdistr(q, meanvalue, distr=c("poisson", "nbinom"), distrcoefs, ...) qdistr(p, meanvalue, distr=c("poisson", "nbinom"), distrcoefs, ...) rdistr(n, meanvalue, distr=c("poisson", "nbinom"), distrcoefs) sddistr(meanvalue, distr=c("poisson", "nbinom"), distrcoefs) ardistr(response, meanvalue, distr=c("poisson", "nbinom"), distrcoefs) checkdistr(distr=c("poisson", "nbinom"), distrcoefs) ```

## Arguments

 `x` vector of (non-negative integer) quantiles. `q` vector of quantiles. `p` vector of probabilities. `n` positive integer value giving the number of random values to return. `response` vector of true observations for calculation of residuals. `meanvalue` non-negative numeric vector of means. `distr` character value giving the distribution. Possible values are currently `"poisson"` (the default) for the `Poisson` distribution and `"nbinom"` for the `NegBinomial` distribution. `distrcoefs` vector of additional distribution coefficients. For the Poisson distribution this argument can be omitted. For the negative binomial distribution it needs to be a vector of length one giving the value for the parameter `size` as specified in `NegBinomial`. `...` additional arguments `logt`, `lower.tail` or `log.p` to be passed to other functions (see `Poisson` or `NegBinomial`).

## Details

Basically, these function are wrappers for specific functions for the respective distribution. The function `ddistr` gives the density of the specified distribution, `pdistr` the distribution function, `qdistr` the quantile function and `rdistr` generates random deviates from this distribution. These functions are a generalisation of the respective functions where `distr` is replaced by either `pois` or `nbinom`. The function `sddistr` returns the standard deviation of the specified distribution. The function `ardistr` calculates Anscombe residuals for given values of the response. The function `checkdistr` is for verification of the arguments `distr` and `distrcoefs`.

## Author(s)

Tobias Liboschik

`Poisson` for the Poisson distribution and `NegBinomial` for the negative binomial distribution.
`tsglm` for fitting a more genereal GLM for time series of counts.