作者: Seung-Ho Kang
DOI: 10.1109/ICITCS.2015.7292916
关键词:
摘要: The performance of network intrusion detection systems based on machine learning techniques largely depends the selected features. However, choosing optimal subset features from a given feature set requires extensive computing resources. To tackle this problem we propose an selection algorithm local search algorithm. In order to evaluate our proposed algorithm, comparisons with composed all 41 are carried out over NSL-KDD data using multi-layer perceptron.