In Rkabacoff/qacStats: Statistical functions for QAC courses

knitr::opts_chunk$set(echo = FALSE, warning = FALSE, message = FALSE)
library(multi_reg)

While base R's lm returns a very informative and useful multiple regression model, these functions produce dull tables and graphs and bury valuable regression information in lists. Developed to refine the lm functions, the multi functions (multi_reg and multi_plot) output the same information returned by the lm functions, but present it in a much more aesthetically pleasing and clear way.

Regression Results

While lm.summary presents regression results through a basic table, multi_reg depicts the results of the linear model through a visually appealing and highly comprehensible stargazer table.

data(mtcars)
regression <- lm(mpg ~ wt + hp + cyl, data = mtcars)
multi_reg(regression, mtcars)

The multi_reg function also enhances the results table embedded in lm.summary by adding a 95% confidence interval for each coefficient as well as stars to represent significant predictors.

a <- multi_reg(regression, mtcars)
a[["coefficients"]]

In order to facilitate user demands to immediately save the results of the multiple regression, the multi_reg function includes the export_data parameter. When export_data = TRUE, this function will save the results of the multiple regression model into an Excel file. The file_name parameter allows the user to customize the name of this Excel file.

Regression Diagnostic Tests

Multiple linear regressions assume that residuals are normally distributed and that independent variables are not highly correlated with one another. While the lm function does not allow clear opportunities to assess these assumptions, multi_reg offers the user a histogram of the model's results as well as a correlation plot depicting the correlations of each predictor in the model.

multi_reg(regression, mtcars, results = FALSE, 
          hist_of_resid = TRUE, corr_plot = TRUE)

Regression Diagnostic Plots

Although the lm.plot function outputs extremely informative plots assessing the validity of the multiple regression model, these plots are exceptionally dull. In order to more persuasively and transparently convey these results, mulit_plot utilizes ggplot to return the same informaion in a more elegant fashion. Multi_plot also includes a Residuals versus Order plot to evaluate the relationship among the residuals over the duration of data collection.