Improving negative selection algorithm in artificial immune systems for computer virus detection

Nguyễn Văn Trường và cs Tạp chí KHOA HỌC & CÔNG NGHỆ 72(10): 53 - 58

Số hóa bởi Trung tâm Học liệu – Đại học Thái Nguyên http://www.lrc-tnu.edu.vn | 53

IMPROVING NEGATIVE SELECTION ALGORITHM IN ARTIFICIAL IMMUNE

SYSTEMS FOR COMPUTER VIRUS DETECTION

Nguyen Van Truong1

, Pham Dinh Lam2*

1 College of Education –TNU; 2 Board of Information Technology – TNU

ABSTRACT

In Artificial Immune Systems (AIS), negative selection algorithms are used widely. This paper

presents the author's research in improving the negative selection algorithm to increase the

performance of AIS applications for detecting computer virus. Our algorithm’s time complexity is

equal to and its space complexity is less than those mentioned in [7]. Furthermore, these

complexities are irrelevant to the size of detector set used. This new valuable characteristic makes

it especially suitable for AISs having ability to detect viruses 100% accurately even with very

large data space.

Keywords: Artificial immune system, Negative selection algorithm, Intrusion detection system,

self, detector

INTRODUCTION

The idea comes from biology has led to the

appearance of some new research areas such

as: artificial neural networks, genetic

algorithms, etc. AIS is an approach to

artificial intelligence system to solve

problems based on the principles, functions

and operational model of biological immune

system. Like the biological immune system,

AIS has a number of important characteristics

such as resistance to noise, unsupervised

learning, memory, distribution, and self￾organization. AIS is evaluated as a new and

effective soft computing method. AIS can be

applied in many fields such as machine

learning, robotics, learning control,

optimization, etc. It is well known through the

applications in the field of computer security

and information security; especially in

building up the Intrusion Detection Systems

(IDS), that can protect computer systems

against intruders and the destruction of

computer viruses or other malicious software

system. General model of IDS are shown in

Figure 1 [4].

Construction method AIS-based IDS (or AIS –

IDS for short) is considered the most

promising direction by network security

researchers. This is reasonable because



Tel: 0919867172; Email: [email protected]

intrusion prevention first appeared is also the

problem that the human biology immune

systems need to be addressed. In fact, the

human biology immune system solves problem

of fighting against the intrusion of bacteria,

virus (or antigen for common) very effectively.

Figure 1. General model of IDS

Thus, researchers believe that the

understanding and re-simulating the training

mechanisms of antibodies to discriminate

between cells of the body (self) and abnormal

cells of the body (non-self), multi-level and

distributed protection mechanism, mechanism

to recognize and to respond quickly to

historical antigen of human biology immune

system can solve the problem of network

intrusion detection. The result of this study