fit_garma: Fit generalized ARMA-Model

In td-berlin/anomalizer: Implementation of Outlier-Detection-Methods.

Description

This function tries to fit a generalized ARMA-Model with given distribution of the dependent variable taken as link-function. As numeric optimization sometimes failes, the function walks through every distribution specified in given order and stops when estimation was successful. Optionally, a matrix of regressor variables can be given. Mandatory, the ARMA order must be given as a 2-length-vector (AR, MA)

Usage

fit_garma(y, order, dists, x.matrix = NA)

Arguments

y	Dependent Variable
order	vector containing AR und MA order
dists	vector of gamlss.family-distributions of y taken as link function
x.matrix	optional matrix of regressors, use to model seasonality

Value

An garmaFit-object

td-berlin/anomalizer documentation built on Feb. 21, 2020, 2:03 a.m.