چکیده (انگلیسی):

Intrusion Detection System (IDS) is one of the component that take part in the system defence, to identify
abnormal activities happening in the computer system. Nowadays, IDS facing composite demands to defeat
modern attack activities from damaging the computer systems. Anomaly-Based IDS examines ongoing traffic,
activity, transactions and behavior in order to identify intrusions by detecting anomalies. These technique
identifies activities which degenerates from the normal behaviours. In recent years, data mining approach for
intrusion detection have been advised and used. The approach such as Genetic Algorithms , Support Vector
Machines, Neural Networks as well as clustering has resulted in high accuracy and good detection rates but with
moderate false alarm on novel attacks. Many researchers also have proposed hybrid data mining techniques. The
previous resechers has intoduced the combination of Fuzzy Clustering and Artificial Neural Network. However,
it was tested only on randomn selection of KDDCup 1999 dataset. In this study the framework experiment
introduced, has been used over the NSL dataset to test the stability and reliability of the technique. The result of
precision, recall and f-value rate is compared with previous experiment. Both dataset covers four types of main
attacks, which are Derial of Services (DoS), User to Root (U2R), Remote to Local (R2L) and Probe. Results had
guarenteed that the hybrid approach performed better detection especially for low frequent over NSL datataset
compared to original KDD dataset, due to the removal of redundancy and uncomplete elements in the original
dataset. This electronic document is a “live” template. The various components of your paper [title, text, tables,
figures and references] are already defined on the style sheet, as illustrated by the portions given in this document.