A novel network security algorithm based on improved support vector machine from smart city perspective

Abstract

Computer generated security concerns have become more modern and complex. Intrusion detection(ID) is a practical issue in the field of computer security whose primary objective is to detect rare attack or assaults and to ensure the security of interior systems. This paper also proposes a semi-class intrusion detection method that combines multiple classifiers to arrange exceptions and typical exercises in a computer system. The abuse detection model is constructed in the light of the decision tree learning-iterative dichotomise 3(DTL-ID3) and is assembled by utilizing the gathered data based on anomaly detection model executed by one class-support vector machine(OC-SVM). In recent years, people have paid more attention to ID/intrusion prevention system (IDS / IPS), which is closely related to the protection and utilization of system management. A few machine-learning standards including neural system, direct hereditary programming, and advanced support vector machines(ASVMs), Bayesian system, multivariate versatile relapse splines, fluffy derivation systems(FIS) and other analogical systems have been researched for the outline of intrusion detection system. In this paper, we build up an amalgam method based on DTL-ID3 and OC-SVM(A-DT and SVM) and evaluate the performance of the projected methodology by using a specific dataset and a crossover method in order to enhance the accuracy of IDS/IPS when contrasted with a singular support vector machine.

Introduction

The fast advancement and improvement of the Internet has brought security problems to systems which is progressively becoming a extraordinary issue and has been a concentration in the ebb and flow exploration. In recent years, people pay more attention to the problem of IDS, which is closely related to the covert use of system management [3]. In any case, it is difficult to detect the assault and the typical system access. In today's IDS, large-scale information grouping and scheduling has become increasingly important and has become a test area. Albeit different apparatuses are projected, they are productive for certain applications adequately, which are used for exponential developing high dimensional information inputs [7], [9]. Intrusion detection systems are designed to protect computer systems from various digital attacks and infections [13]. The intrusion detection system constructs a robust feature model and examples to identify the general practices of system information described by nonstandard practices. Two basic hypotheses in intrusion detection are studied, for example, client and program exercises can be recognized by PC systems according to system reviewing mechanisms, and ordinary and intrusion exercises must have particular practices. The field of intrusion detection consists of two different approaches, that are abuse detection and anomaly detection [17], [18]. The basic idea of misuse of investigation is to detect the attack of a certain type or target in some way, and even identify the types of these attacks. In view of these signs, this method identifies attacks by describing the criteria for each known attack [1]. The trouble for identifying obscure assaults has become a fundamental drawback in the mark-based method. The primary objective of the anomaly detection method is to describe the typical activities of the manufacturing factual model. In this point, any deviation from this model can be viewed as an anomaly, and perceived as an assault [20]. When this approach is utilized, it can identify obscure assaults hypothetically, despite the fact happened now and again, the considered approach gives rise to high false assault rate. Given the general manufacture models in the past few years, people are keen to develop new manufacturing models [6], [10].

Anomaly detection approach is one of the extremely dynamic researches in the machine learning group, which has been the theme of presented numerous articles over many years. The best approach depends on gathering information from typical operations of the system. In view of this information portraying ordinariness, if any deviation is seen in any case, it would be considered as an anomaly [11]. A few machine learning standards include the hidden Markov model, bolster VM, fake neural system, counterfeit neural system and multivariate versatile relapse splines fluffy surmising systems, which have been researched for the outline of IDS [19]. In the manuscript, we conduct researches and assess the performance of OC-SVM. The proposed amalgam method based on decision tree learning-ID3 and OC-SVM is a combination of A-DT and SVM. Compared with the different methods, it can improve the accuracy of IDS intrusion detection system by using half method [27]. The rest of the paper is organized as the follows. In the Section 2, we review the state-of-the-art related works; in the Section 3, we introduced IEEE Transactions on Reliability; in the Section 4, we discussed the DTL-ID3; in the Section 5, we analysed the A-DT & SVM; in the Section 6, we implemented the proposed method with the experimental simulation; the Section 7 summarized the work.