A Highly Efficient Intrusion Detection Method Based on Hierarchical Extreme Learning Machine

Abstract

Cyber security is becoming more and more concerned by people nowadays. Intrusion detection systems (IDSs) is a major approach to ensure the confidentiality, integrity and availability of network system resources. There are many machine learning techniques applied to IDSs. In this paper, we propose a novel and rapid technique based on Hierarchical Extreme Learning Machine (H-ELM) for intrusion detection. We use NSL-KDD 2009 dataset to evaluate our method. Comparing our method with other widely used machine learning methods such as k-Nearest Neighbor (k-NN), Random Forest (RF) and Extreme Learning Machine (ELM), the experimental results show that H-ELM can perform better than or similar to other methods in overall accuracy of 72.87%, while only spends a total time of 2.04 s which is much faster than other methods.