Analysis of data pre-processing influence on intrusion detection using NSL-KDD dataset

Data pre-processing for machine learning methods is key step for knowledge discovery process. Depending on nature of the data, pre-processing might take the majority time of data analysis. Correctly prepared data for processing guarantees precise and reliable results of data analysis. This paper analyses initial data pre-processing influence to attack detection accuracy by using Decision Trees, Naïve Bayes and Rule-Based classifiers with NSL-KDD dataset. In addition, the results of detected attacks accuracy dependency by selecting different attacks grouping options and using ensembles of various classifiers are presented.