| Type: | Package |
| Title: | Automated Linear Regression Diagnostic |
| Version: | 0.10 |
| Description: | Provides a set of streamlined functions that allow easy generation of linear regression diagnostic plots necessarily for checking linear model assumptions. This package is meant for easy scheming of linear regression diagnostics, while preserving merits of "The Grammar of Graphics" as implemented in 'ggplot2'. See the 'ggplot2' website for more information regarding the specific capability of graphics. |
| Depends: | R (≥ 3.2.2) |
| License: | MIT + file LICENSE |
| Encoding: | UTF-8 |
| RoxygenNote: | 7.2.3 |
| Imports: | MASS, ggplot2, gridExtra |
| URL: | https://github.com/yeukyul/lindia |
| BugReports: | https://github.com/yeukyul/lindia/issues |
| NeedsCompilation: | no |
| Packaged: | 2023-08-10 22:22:59 UTC; tlee |
| Author: | Yeuk Yu Lee [aut, cre], Samuel Ventura [aut] |
| Maintainer: | Yeuk Yu Lee <lyeukyu@gmail.com> |
| Repository: | CRAN |
| Date/Publication: | 2023-08-10 22:40:02 UTC |
Plot boxcox graph in ggplot with suggested lambda transformation
Description
Plot boxcox graph in ggplot with suggested lambda transformation
Usage
gg_boxcox(fitted.lm, showlambda = TRUE, lambdaSF = 3, scale.factor = 0.5)
Arguments
fitted.lm |
a fitted linear model (i.e. lm, glm) that contains fitted regression |
showlambda |
logical; controls whether lambda value should be displayed on graph. Defaults to TRUE |
lambdaSF |
numeric; controls to how many significant figure is lambda rounded to. Defaults to 3. |
scale.factor |
numeric; scales the point size and linewidth to allow customized viewing. Defaults to 0.5. |
Value
A ggplot object that contains boxcox graph
Examples
library(MASS)
data(Cars93)
cars_lm <- lm(Price ~ Passengers + Length + RPM, data = Cars93)
gg_boxcox(cars_lm)
Plot cook's distance graph
Description
Plot cook's distance graph
Usage
gg_cooksd(
fitted.lm,
label = TRUE,
show.threshold = TRUE,
threshold = "convention",
scale.factor = 0.5
)
Arguments
fitted.lm |
a fitted linear model (i.e. lm, glm) that contains fitted regression |
label |
logical; whether or not to label observation number larger than threshold. Default to TRUE. |
show.threshold |
logical; determine whether or not threshold line is to be shown. Default to TRUE. |
threshold |
string; determining the cut off label of cook's distance. Choices are "baseR" (0.5 and 1), "matlab" (mean(cooksd)*3), and "convention" (4/n and 1). Default to "convention". |
scale.factor |
numeric; scales the point size and linewidth to allow customized viewing. Defaults to 0.5. |
Value
A ggplot object that contains a cook's distance plot
Examples
library(MASS)
data(Cars93)
cars_lm <- lm(Price ~ Passengers + Length + RPM, data = Cars93)
gg_cooksd(cars_lm)
Plot all diagnostic plots given fitted linear regression line.
Description
Plot all diagnostic plots given fitted linear regression line.
Usage
gg_diagnose(
fitted.lm,
theme = NULL,
ncol = NA,
plot.all = TRUE,
mode = "all",
scale.factor = 0.5,
boxcox = FALSE,
max.per.page = NA
)
Arguments
fitted.lm |
lm object that contains fitted regression |
theme |
ggplot graphing style using 'ggplot::theme()'. A ggplot graphing style to apply to all plots. Default to null. |
ncol |
specify number of columns in resulting plot per page. Default to make a square matrix of the output. |
plot.all |
logical; determine whether plot will be returned as an arranged grid. When set to false, the function will return a list of diagnostic plots. Parameter defaults to TRUE. |
mode |
A string. Specifies which set of diagnostic plots to return: * 'all' (the default) * 'base_r': only graphs included in the base R 'plot(lm(...))' (i.e. residual vs fitted, QQ plot, scale location, residual vs leverage) |
scale.factor |
numeric; scales the point size, linewidth, labels in all diagnostic plots to allow optimal viewing. Defaults to 0.5. |
boxcox |
logical; detemine whether boxcox plot will be included. Parameter defaults to FALSE. |
max.per.page |
numeric; maximum number of plots allowed in one page. |
Value
An arranged grid of linear model diagnostics plots in ggplot. If plot.all is set to FALSE, a list of ggplot objects will be returned instead. Name of the plots are set to respective variable names.
Examples
library(MASS)
data(Cars93)
# a regression with categorical variable
cars_lm <- lm(Price ~ Passengers + Length + RPM + Origin, data = Cars93)
gg_diagnose(cars_lm)
# customize which diagnostic plot is included
plots <- gg_diagnose(cars_lm, plot.all = FALSE)
names(plots) # get name of the plots
exclude_plots <- plots[-c(1, 3) ] #exclude certain diagnostics plots
include_plots <- plots[c(1, 3)] # include certain diagnostics plots
plot_all(exclude_plots) # make use of plot_all() in lindia
plot_all(include_plots)
Plot quantile-quantile plot (QQPlot) in ggplot with qqline shown.
Description
Plot quantile-quantile plot (QQPlot) in ggplot with qqline shown.
Usage
gg_qqplot(fitted.lm, scale.factor = 1)
Arguments
fitted.lm |
a fitted linear model (i.e. lm, glm) that contains fitted regression |
scale.factor |
numeric; scales the point size and linewidth to allow customized viewing. Defaults to 1. |
Value
A qqplot with fitted qqline
Examples
library(MASS)
data(Cars93)
cars_lm <- lm(Price ~ Passengers + Length + RPM, data = Cars93)
gg_qqplot(cars_lm)
Generate residual plot of residuals against predictors
Description
Generate residual plot of residuals against predictors
Usage
gg_resX(
fitted.lm,
plot.all = TRUE,
scale.factor = 0.5,
max.per.page = NA,
ncol = NA
)
Arguments
fitted.lm |
a fitted linear model (i.e. lm, glm) that contains fitted regression |
plot.all |
boolean value to determine whether plot will be return as a plot arranged using 'grid.arrange()'. When set to false, the function would return a list of residual plots. Parameter defaults to TRUE. |
scale.factor |
numeric; scales the point size and linewidth to allow customized viewing. Defaults to 0.5. |
max.per.page |
numeric; maximum number of plots allowed in one page. Parameter defaults to fit all plots on one page. |
ncol |
specify number of columns in resulting plot per page. Default to make a square matrix of the output. |
Value
An arranged grid of residuals against predictor values plots in ggplot. If plotall is set to FALSE, a list of ggplot objects will be returned instead. Name of the plots are set to respective variable names.
Examples
library(MASS)
data(Cars93)
# a regression with categorical variable
cars_lm <- lm(Price ~ Passengers + Length + RPM + Origin, data = Cars93)
gg_resX(cars_lm)
# customize which diagnostic plot is included by have gg_resX to return a list of plots
plots <- gg_resX(cars_lm, plot.all = FALSE)
names(plots) # get name of the plots
exclude_plots <- plots[-1 ] #exclude certain residual plots
include_plots <- plots[1] # include certain residual plots
plot_all(exclude_plots) # make use of plot_all() in lindia
plot_all(include_plots)
Generate residual plot of residuals against fitted value
Description
Generate residual plot of residuals against fitted value
Usage
gg_resfitted(fitted.lm, scale.factor = 1)
Arguments
fitted.lm |
a fitted linear model (i.e. lm, glm) that contains fitted regression |
scale.factor |
numeric; scales the point size and linewidth to allow customized viewing. Defaults to 1. |
Value
A ggplot object
Examples
library(MASS)
data(Cars93)
cars_lm <- lm(Price ~ Passengers + Length + RPM, data = Cars93)
gg_resfitted(cars_lm)
Generate histogram of residuals in ggplot.
Description
Generate histogram of residuals in ggplot.
Usage
gg_reshist(fitted.lm, bins = NULL)
Arguments
fitted.lm |
a fitted linear model (i.e. lm, glm) that contains fitted regression |
bins |
bin size for histogram |
Value
A ggplot object
Examples
library(MASS)
data(Cars93)
cars_lm <- lm(Price~ Passengers + Length + RPM, data = Cars93)
gg_reshist(cars_lm)
# specify number of bins
gg_reshist(cars_lm, bins = 20)
Plot residual versus leverage plot in ggplot.
Description
Plot residual versus leverage plot in ggplot.
Usage
gg_resleverage(fitted.lm, method = "loess", se = FALSE, scale.factor = 1)
Arguments
fitted.lm |
a fitted linear model (i.e. lm, glm) that contains fitted regression |
method |
smoothing method of fitted line on scale-location plot. eg. "lm", "glm", "gam", "loess", "rlm". See https://ggplot2.tidyverse.org/reference/geom_smooth.html for more details. |
se |
logical; determines whether se belt should be plotted on plot |
scale.factor |
numeric; scales the point size and linewidth to allow customized viewing. Defaults to 1. |
Value
A ggplot object that contains residual vs. leverage graph
Examples
library(MASS)
data(Cars93)
cars_lm <- lm(Price ~ Passengers + Length + RPM, data = Cars93)
gg_resleverage(cars_lm)
Plot scale-location (also called spread-location plot) in ggplot.
Description
Plot scale-location (also called spread-location plot) in ggplot.
Usage
gg_scalelocation(fitted.lm, method = "loess", scale.factor = 1, se = FALSE)
Arguments
fitted.lm |
a fitted linear model (i.e. lm, glm) that contains fitted regression |
method |
smoothing method of fitted line on scale-location plot. eg. "lm", "glm", "gam", "loess", "rlm". See https://ggplot2.tidyverse.org/reference/geom_smooth.html for more details. |
scale.factor |
numeric; scales the point size and linewidth to allow customized viewing. Defaults to 1. |
se |
logical; determines whether se belt should be plotted on plot |
Value
A ggplot object that contains scale-location graph
Examples
library(MASS)
data(Cars93)
cars_lm <- lm(Price ~ Passengers + Length + RPM, data = Cars93)
gg_scalelocation(cars_lm)
Plot all given plots in a square matrix form.
Description
Plot all given plots in a square matrix form.
Usage
plot_all(plots, ncol = NA, max.per.page = NA)
Arguments
plots |
a list of plots |
ncol |
numeric; the number of column that the arranged grid need to be. defaults to fitting all plots in square matrix |
max.per.page |
numeric; maximum number of plots allowed in one page. |
Value
plots in a given list arrangeed using gridExtra
Examples
library(MASS)
data(Cars93)
# a regression with categorical variable
cars_lm <- lm(Price ~ Passengers + Length + RPM + Origin, data = Cars93)
plots <- gg_diagnose(cars_lm, plot.all = FALSE)
names(plots)
selected.plots <- plots[-c(2, 5)]
plot_all(selected.plots)