spam classification using logistic regression
Used email spam data set Data preprocessed three ways: standardize columns so they all have mean 0 and unit variance, transformation of features using log(xij+ 0.1), and binarizing featues using I(xij> 0)