# prior.nl: Prior parameter distribution for the NL model In lbelzile/BMAmevt: Multivariate Extremes: Bayesian Estimation of the Spectral Measure

## Description

Density and generating function of the prior distribution.

## Usage

 `1` ```prior.nl(type = c("r", "d"), n, par, Hpar, log, dimData = 3) ```

## Arguments

 `type` One of the character strings `"r"`, `"d"` `n` The number of parameters to be generated. Only used if `type == "r"`. `par` A vector of length four, with component comprised between 0 and 1 (both end points excluded for the first element and 1 included for the others): The parameter where the density is to be taken. Only used if `type=="d"`. In the NL model, `par` is of length 4. The first element is the global dependence parameter, the others are partial dependence parameter between pairs (12), (13), (23) respectively. In the NL model, `par` is of length 4. The first element has the same interpretation as in the NL model, the subsequent ones are dependence parameters between `Hpar` list of Hyper-parameters : see `nl.Hpar` for a template. `log` logical. Should the density be returned on the log scale ? Only used if `type=="d"` `dimData` The dimension of the sample space, equal to 3. Only for compatibility with e.g. `posteriorMCMC`.

## Details

The four parameters are independent, the logit-transformed parameters follow a normal distribution.

## Value

Either a matrix with `n` rows containing a random parameter sample generated under the prior (if type == "d"), or the (log)-density of the parameter `par`.

Anne Sabourin

## Examples

 ```1 2 3 4``` ```## Not run: prior.nl(type="r", n=5 ,Hpar=get("nl.Hpar")) ## Not run: prior.trinl(type="r", n=5 ,Hpar=get("nl.Hpar")) ## Not run: prior.pb(type="d", par=rep(0.5,2), Hpar=get("nl.Hpar")) ```

lbelzile/BMAmevt documentation built on June 13, 2019, 12:43 p.m.