I have a dataset of text messages of which I'm trying to filter out the spam from the legitimate ones. I have roughly 4600 pieces of data spread among 57 features and then their classification as spam or not. I have four 'versions' of the data, one the regular data and the other three I have applied various types of preprocessing.
I'm supposed to be fitting each of these to a regression model using ridge regressing, and must use cross validation to calculate the ridge regression parameter. I have a loose understanding of what cross validation is but I'm confused as to how I apply it to this situation, particularly how to split my data. Could I get some guidelines/pointers for this?