Research Information System
Thesis Type: Postgraduate
Institution Of The Thesis: Gazi Üniversitesi, Fen Bilimleri Enstitüsü, Turkey
Approval Date: 2016
Student: ÇETİN KAYA
Supervisor: OKTAY YILDIZ
Open Archive Collection: AVESIS Open Access Collection
Abstract:Internet is an indispensable part of our daily lives. Increasing web applications and number of users has brought some risks in terms of data security. An intrusion detection system is one of the important tools for network security and is used successfully to detect attacks and unexpected demands made to secure access to internal network. Today, many researchers are working in order to realize more effective intrusion detection system. For this purpose, there are many intrusion detection system was performed using different machine learning techniques in the literature but according to the types of attacks, which machine learning techniques are more successful in IDS. It does not answer this question. In our study, with the experiments performed, the most commonly used machine learning techniques in intrusion detection systems that Bayesian network, support vector machines, decision trees, neural networks and k nearest neighbor algorithm performance analysis was conducted. According to the types of attacks, accuracy, specificity, sensitivity, accuracy, F-Measure values were examined and determined the most successful classifiers. With this study, Intrusion detection with machine learning techniques for future studies aimed to gain a perspective on. KDD CUP99 and NSL-KDD datasets were used in experimental studies.