An Intelligent Data Mining Approach to Detect Intrusions in Social Programming Networks

Author:   Hayder Hussein Azeez* and Mustafa Musa Jaber

Download PDF

Abstract: Data mining it’s one of the successful techniques for intrusion detection in the fast growing internet world. Moreover, they are extensively used to develop intrusion detection systems that classify the users based on the behavior in various contexts. Soft computing techniques are used in data mining for making effective decisions by the applications of neural networks, fuzzy logics and genetic algorithms. Moreover, Fuzzy Rule-Based Classification Systems (FRBCSs) are productive and they are recognized well in the machine learning area, since they can present an interpretable decision model for the end user. The use of fuzzy logic algorithms helps to develop an efficient classification model that can detect the network intrusions more effectively. In this work, new feature selection algorithms namely Fuzzy based intelligent non-dominated particle sorting optimization (F-INDPS) are proposed to select The optimal number of features from dataset. These features contribute in classification process for forming rules and to apply them for effective decision making. An experimental result of the proposed system increases the security through the intrusion detection system and Total completion time has been reduced. Moreover, the proposed system increased the throughput and overall performance of the networks by enhancing the packet delivery ratio and by reducing the delay in communication and thereby provided more reliable and secured communication in social programming networks.