Comparison of logistic regression with random forest