AN EMPIRICAL EVALUATION FOR THE INTRUSION DETECTION FEATURES BASED ON MACHINE LEARNING AND FEATURE SELECTION METHODS