Feature Selection Using Euclidean Distance and Cosine Similarity for Intrusion Detection Model

Nowadays, data mining plays an important role in many sciences, including intrusion detection system (IDS). However, one of the essential steps of data mining is feature selection, because feature selection can help improve the efficiency of prediction rate. The previous researches, selecting features in the raw data, are difficult to implement. This paper proposes feature selection based on Euclidean Distance and Cosine Similarity which ease to implement. The experiment results show that the proposed approach can select a robust feature subset to build models for detecting known and unknown attack patterns of computer network connections. This proposed approach can improve the performance of a true positive intrusion detection rate.

I. Introduction

The Internet and local area networks are growing larger in recent years. As about 1,463,632,361 people all over the world use the Internet, they are unconsciously encountering the number of security threats such as viruses, worms and attacks from hackers [1], [2]. Now firewalls, anti-virus software, message encryption, secured network protocols, password protection and so on are not sufficient to guarantee the security in computer networking, which some intrusions take advantages of weaknesses in computer systems. Therefore, intrusion detection is becoming a more and more important technology which monitors network traffic and identifies network intrusion such as anomalous network behaviors, unauthorized network access, and malicious attacks to computer systems [3].