regression.diagnostic.cmp: CMP Regression Diagnostic

In mpcmp: Mean-Parametrized Conway-Maxwell Poisson (COM-Poisson) Regression

Description

This suite of functions provides the basic quantities which are used in forming a wide variety of diagnostics for checking the quality of regression fits.

Usage

## S3 method for class 'cmp'
influence(model, ...)

## S3 method for class 'cmp'
hatvalues(model, ...)

## S3 method for class 'cmp'
rstandard(
  model,
  infl = influence.cmp(model),
  type = c("deviance", "pearson"),
  ...
)

## S3 method for class 'cmp'
cooks.distance(
  model,
  infl = influence(model),
  res = infl$pear_res,
  hat = infl$hat,
  ...
)

Arguments

model	an object class 'cmp', obtained from a call to glm.cmp.
...	other arguments passed to or from other methods (currently unused).
infl	influence structure as returned by influence, only for rstudent and cooks.distance.
type	type of residuals for rstandard. The alternatives are: 'deviance' (default), and 'pearson'.
res	residuals, with proper default.
hat	hat values H[i,i], see default.

Examples

data(attendance)
M.attendance <- glm.cmp(daysabs~ gender+math+prog, data=attendance)
influence(M.attendance)
hatvalues(M.attendance)
rstandard(M.attendance, type = "pearson")
cooks.distance(M.attendance)

mpcmp documentation built on Oct. 26, 2020, 9:07 a.m.