cooksd: Evaluates Cook's Distance.

In roland-hochmuth/gamlssdiag: GAMLSS Diagnostics

Description

Evaluates Cook's Distance.

Usage

cooksd(fn, formula, data, family)

Arguments

fn	The function to use when creating the model. Either "glm" or "gamlss".
formula	The formula to use in the model.
data	The dataframe.
family	The distribution family. E.g. "poisson" if using glm. PO or NBI if using gamlss.

Value

A vector of distances

Note

Although lm/glm stores the original data in the model at m$data, that is not the case in GAMLSS. Therefore, unlike the function cooks.distance which only requires a model parameter, this function requires the formula and data.

References

https://www.ime.usp.br/~abe/lista/pdf1USQwcGBX1.pdf

Robust Diagnostic Regression Analysis, Anthony Atkinson and Marco Riani

Examples

cooksd(glm, f, df, "poisson")
cooksd(gamlss, f, df, PO)
cooksd(gamlss, f, df, NBI)

roland-hochmuth/gamlssdiag documentation built on May 20, 2019, 3:01 p.m.