Web Anomaly Host Based IDS, a Machine Learning Approach

Please enable javascript in your browser.

Khalkhali, Iman | 2011

683 Viewed

Type of Document: M.Sc. Thesis
Language: English
Document No: 41615 (52)
University: Sharif University of Technology, International Campus, Kish Island
Department: Science and Engineering
Advisor(s): Azmi, Reza; Khansari, Mohammad
Abstract:
Web servers and web applications are susceptible to different attacks. In order to detect web-based attacks Intrusion detection systems (IDS) should be equipped with a large number of signatures. Unfortunately various types of web threats are increasingly growing and so detection and prevention of all these new and old attacks is exhaustive and really difficult.This thesis represents a designed system for intrusion detection that uses different techniques to discover vulnerabilities with derived patterns and also some user behavior based attacks against web applications. This was done by using new dataset which was generated by new log file.The primary objective of this thesis shows the design and customable structure for web hostbased anomaly intrusion detection and also effectiveness of more accurate log files, lead to high performance of anomaly detection techniques. Finally experimental results show that high
detection rate and small amount of false alarms generated with new dataset and improved detection techniques
Keywords:
Data Mining ; Machine Learning ; Neural Network ; Radial Basis Function ; Anomaly ; Web Based Intrusion Detection System ; Log File

No TOC