gev_fit: Fit a GEV model
In dannyjameswilliams/danielR: Collection of Danny's R Code

gev_fit

R Documentation

Fit a GEV model

Description

gev_fit is used to fit a Generalised Extreme Value (GEV) linear model. This is a wrapper function, using gev_likfit and gev_grad to obtain parameter estimates via maximum likelihood.

Usage

gev_fit(
  formulas,
  data,
  response,
  rl_n = 100,
  links = list("identity", "exponential", "identity")
)

Arguments

`formulas`	a list with three elements, all objects of class "`formula`", for the location, scale and shape parameter in that order. A symbolic description of the model
`data`	data frame containing all variables listed in `formulas`
`response`	the observed maxima corresponding to those variables in `data`
`rl_n`	the value of n of which to calculate the return level of period n, if `NULL`, the return level is not calculated
`links`	a list with three elements, all strings detailing which link function to use for the location, scale and shape parameter in that order

Details

The formulas object have formulas specified symbolically, of the form ~ predictor1 + predictor2 + ... where both predictor1 and predictor2 are variables in data.

The GEV model is fit by using optim with the BFGS method.

Value

an S3 object of class "gev", containing

`loglik`	value of the final log-likelihood at the values of `par`
`par`	list of maximum likelihood parameter estimates for each GEV distribution parameter
`par_se`	standard errors of the parameters in `par`
`modX`	list of model matrices used for each parameter
`link`	list of link functions inputted
`formula`	list of formulae inputted
`big_formula`	combination of unique elements in all formulae used
`data`	data frame inputted
`response`	response vector (observed maxima) inputted