从所有回归变量组合中可视化所有可能的线性回归预测

Question

So I have a set of 8 variables that I'm regressing against a dependent variable. Obviously I want to avoid over fitting so I want to see what all combinations of my 8 variables look like from a visual perspective to see if their is a correlation around the true model. Is there an R function or package that can help me do this? So far I've been using the "MuMin" package and the dredge() function to run all combinations of variables. What I'd like to do now is visualize all of them too. Any recommendations are appreciated.

Answer 1

What you are trying to do is called a Pairs Plot. Here is how you can implement it:

library(GGally)
library(ggplot2)

ggpairs(data = your_data,
        columns = c('column name 1', 'column name 2', 'column name 3'),
        title = 'Pairs Plot')

The columns option within the ggpairs function allows you to only choose specific columns. If you want to automatically use every column in the pairs plot, omit the columns option:

ggpairs(data=your_data, title='Pairs Plot')

This is an example of a Pairs Plot: