partial_plot: Partial Regression Plots for visualizing relationship between...
In mattmills49/modeler: Exploratory Data Analysis and Modeling Helper Functions

Description Usage Arguments Value Methods (by class) Examples

partial_plot accepts a fitted regression object and the name of the variable you wish to view the partial regression plot of as a character string. It returns a ggplot object showing the independent variable values on the x-axis with the resulting predictions from the independent variable's values and coefficients on the y-axis. This shows the relationship that the model has estimated between the independent variable the dependent variable. You can determine which prediction level the plot is against using the response parameter which is a logical value defaulted to FALSE.

partial_plot(fitted_model, ...)

## S3 method for class 'gam'
partial_plot(fitted_model, variable, response = F, rug = F)

`fitted_model`	a complete regression model object
`variable`	the name of the independent variable as a character string
`response`	logical indicating if the plot should be on the linear prediction scale or the response scale. Defaults to `FALSE`
`rug`	logical indicating if a rug plot should be added on the independent variable axis

a ggplot2 object of the partial regression plot

gam: provides partial plots for fitted gam models

library(mgcv)
car_gam <- gam(mpg ~ s(hp), data = mtcars)
partial_plot(car_gam, "hp")

# Response level changes look for non-gaussian families
am_gam <- gam(factor(am) ~ s(hp), data = mtcars, family = "binomial")
partial_plot(am_gam, "hp") # on the log odds scale
partial_plot(am_gam, "hp", response = T) # on the probability scale