Identifying relevant predictors and interactions

I have a data set comprising of 12,000 observations and 1,100 potential predictors (continuous and nominal) for a binary outcome/target variable.

The research question is to identify interesting predictors and interactions for the binary outcome/target variable.

Given the large number of potential predictors, I would like to exclude irrelevant predictors before modelling the data. Can somebody please recommend methods for feature selection that take interactions into account?
Last edited: