outlier_influence: outlier_influence

In Jiaxin-Cara-Qian/testpackage: Resisuals and Outlier Influence in Linear Regression

Description

outlier_influence computes three diagnostic statistics for measuring influential points:

DIFFTS

Difference in the fitted value when ith observation is removed

Cook's Distance

shows the influence of each observation on the fitted response values

COVRATIO

influence on variance-covariance matrix

Usage

outlier_influence(input_data, fit.model,
  option = c("dffits", "cd" , "cvr"),
    high.influence = FALSE)

Arguments

input_data	a data frame containing the variables in th model
fit.model	takes an R object, returned by lm
option	which measure of influence diagnostics should be computed. "dffits" is for DFFITS, "cd" is for Cook's distance, "cvr" is for COVRATIO
high.influence	(Default)FALSE; if TRUE, influential observation along with their respective diagnostic statistics will be returned.

Value

If high.influence is set TRUE, each diagnostic statistic will return a list containing two elements.

• dffits, Cook's Distance, COVRATIO - depends on which option is chosen

• outliers - which observations are considered influential

Examples

##Use demo_data (See ?demo_data for more information about the dataset)
m = lm(fev~A+H+M+S, data = demo_data)
outlier_influence(demo_data, m, option = c("dffits"))
outlier_influence(demo_data, m, option = c("cd"))
outlier_influence(demo_data, m, option = c("cvr"), high.influence = TRUE)

Jiaxin-Cara-Qian/testpackage documentation built on Dec. 18, 2021, 1:30 a.m.