An Intrusion Detection System for Web Server Based on Smooth Support Vector Machine

• Main Authors: Lan Shao-Wei, 藍紹緯
• Other Authors: Horng Shi-Jinn
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/90393999713637569996

碩士 === 國立臺灣科技大學 === 資訊工程系 === 92 === We propose a Web log analysis system based on Smooth Support Vector Machine, using the advantage of machine learning rapidly to improve the system reliability. With the hierarchy of Smooth Support Vector Machine, we can find the category to which the attack belongs and even detect the unknown attacks never found before. The known attacks’ detecting rate is 100% and the unknown attacks’ detecting rate is 96% in our system. The kernel of my Web log analysis system is to utilize the machine leaning Smooth Support Vector Machine to cope with swiftly changing attacks, and it can reduce the training time to protect the security of all kinds of network applications immediately. The system we develop on the platform of Microsoft Windows, many people of our country use the Microsoft Windows production, so our system can benefit more people.