olsdiagnostic: olsdiagnosticR

In Kale-S/isnormalr: olsdiagnosticR

Description

Regression diagnostics for a linar regression model.

Usage

olsdiagnostic(object, normtest = "Shapiro-Wilk", pval = 0.05)

Arguments

object	linear regression model from class 'lm'
normtest	the type of normality test. Possible test are 'Shapiro-Wilk', 'Jarque-Bera' or 'Anderson-Darling'
pval	p-value for the hypothesis test. Default is 0.05

Details

olsdiagnostic was developed for the statistical consulting of the University of Göttingen to assist students in checking common assumptions of linear regression quickly and easily. It also provides students with an automated recommendation for the further procedure if an assumption is not fulfilled.

Value

A olsdiagnosticR class with following entries:

Normality_Tests:

This could be the Shapiro-Wilk test, Jarque-Bera test, Anderson-Darling test or Cramer-von-mises test. Further information can be found in the documentation:

Shapiro_Wilk_test

jarque_bera

anderson_darling_test

Outlier:

Possible.Outlier: a data.frame with possible outlier

Cooks.Distance: a vector with the cook's distance values

Hat.Values a vector with the lavarage values

Studentized.Residuals a vector with the studentized residuals

Standardized.Residuals a vector with the standardized residuals

Further information can for these values be found in the documentation:

influence_observation

Multicollinearity:

VIF: a vector of the variance inflation factors Further information can be found in the documentation:

VIF

Contain.Columns: A string with the columns that would be avoid multicollinearity in the data

Heteroskedasticity:

A Breusch-Pagan 'htest' object.

Further information can be found in the documentation:

bp_test

Plots:

QQ_Plot: qq_plot

Residual_Hist: resid_hist

Box_Plot_X: box_plot_x

Regressor_hist: hist_x

Cooks_Distance_Plot: plot_cd

Bubble_Plot: influence_plot

Spread_Level_Plot: influence_plot

Examples

## Not run: 
e_norm <- rnorm(n)
e_exp <- rexp(n)

x1 = rnorm(n = n, mean = 80, sd = 10)
x2 = rnorm(n = n, mean = 70, sd = 5)
x3 = 2 * x1 + 4 * x2 + rnorm(n, mean = 20, sd = 5)

y_n = 3 + x1 + x2 + e_norm
y_m_e = 4 + x1 + x2 + x3 + e_exp

mod_n <- lm(y_n ~ x1 + x2)
mod_m_e <- lm(y_m_e ~ x1 + x2 + x3)

olsdiagnosticR::olsdiagnostic(mod_n)
olsdiagnosticR::olsdiagnostic(mod_m_e, normtest = 'Jarque-Bera')

## End(Not run)

Kale-S/isnormalr documentation built on Sept. 23, 2019, 5:48 a.m.