Intrusion detection system with decision tree and combine method algorithm

Abstract

By developments in social networks, the number of attacks and intrusions into these networks are
increasing in different ways. Hence an intrusion detection system could play an important role in
protecting the systems against the hackers. In order to protect computer networks and systems, different
solutions introduced which are called intrusion detection system. The purpose of such systems is to
protect the computers and networks from prohibited uses, damages and misuse of such systems against
internal and external intruders and hackers. Each intrusion detection system may use two different
approaches to identify any suspicious (abnormal) activity or intrusion. In order to identify the misuse of
systems, intrusion patterns are used; however those approaches used to identify abnormal activities, the
normal manner of users are used as the pattern. Decision tree is one of the most famous and oldest
methods in data analysis to make categorizing model. In those algorithms based on decision tree, the
output is represented as a tree which is made of different states of values. Since the results of them could
be interpreted and do not need all input parameters, decision trees are in scope their structures are easy to
interpret too. Efficiency of a system strictly depends on selecting the properties. Since by an increase in
properties, the computational cost also increases, it seems that design and implementation of systems
needs the minimum number of them. The proposed method uses a training set of KDD-Cup99. The
proposed method uses three main learning algorithms, SVM, Naive Bayes and J48 decision tree is
implemented and evaluated separately. These algorithms are also implemented and evaluated individually
as well. The results show the superiority of the proposed method with 97% efficiency using J48 learning
algorithm and ADABOOST classification by reducing the dimension IG method.