Handbook for the diagnoser R package

Robert Schnitman
diagnoser / OLS / residuals / diagnostics

The book for my package diagnoser has just been published online! You may view it as a Gitbook (https://rs-diagnoser.netlify.com/) or PDF (https://github.com/robertschnitman/diagnoser/blob/master/docs/diagnoser_handbook.pdf).

I recommend reading it if you are interested in a new perspective on diagnosing your model residuals!

Installing diagnoser

The library diagnoser currently is only installable via GitHub and is contingent on R versions at or above 3.4.2. To install the package, first install devtools so that you may make use of the function install_github, referencing diagnoser by the package creator’s username (“robertschnitman”) followed by “/diagnoser” as shown in the code below:

## Ensure that you are running R 3.4.2 or higher.
## Package Dependencies:
#     lazyeval (>= 0.2.1)
#  Package Imports:
#     ggplot2 (>= 2.2.1), 
#     gridExtra (>= 2.3), 
#     scales (>= 0.5.0), 
#     car (>= 2.1)

# Install library necessary for installing diagnoser.
install.packages("devtools") 

# Install diagnoser via devtools.
devtools::install_github("robertschnitman/diagnoser")

Example 1: diagnose()

model.lm <- lm(data = mtcars, formula = mpg ~ wt + gear)

diagnose(model.lm, 
         fit_type      = 'response', 
         residual_type = 'response')

Example 2: ggdiagnose()

ggdiagnose(model.lm, 
           fit_type      = 'response', 
           residual_type = 'response',
           freqpct       = TRUE, 
           alpha         = 0.5)

Example 3: cdiagnose()

model.lm <- lm(data = Orange, formula = log(circumference) ~ age)

cdiagnose(model.lm, 
          fit_type      = 'response',
          residual_type = 'response',
          se            = FALSE,
          alpha         = 1)

See these functions and more from my book!