A comparative study of Feature Selection techniques for Intrusion Detection

Feature Selection plays an important role in Intrusion Detection, where a large number of features extracted from whole data needs to be analyzed. Feature relevance is the basic measurement in feature selection techniques. In this paper, different feature selection techniques are analyzed. By using pre-processed data set, various feature selection techniques are compared. The NSL - KDD dataset is used for the evaluation purpose. Various Feature Selection techniques are applied to NSL-KDD data set for reduced training & test data sets. Naive Bayes Classifier is used to classify in this. We have compared all the experimented results by using different performance metrics like TP rate, FP rate, Precision, ROC area, Kappa Statistic and Classification Accuracy.