I have a dataset consisting of a lot of categorical variables.
I want to perform a regression model with them. R automatically transforms them in binary variables in the model output, which is great.
Before constructing the model however, I would like to analyse their pairwise correlations.
So, what is the best way to do this? Transform categorical variables to binary and calculating the correlation matrix for all variables?
Also, what is the best way to select the most suitable variables for the regression model in the next step?
Thanks for any advise!