Snatching Masqueraders from Web Server Logs

Sungdeok Cha: Computer Science Division, EECS Department, Korea Advanced Institute of Science and Technology, Daejeon, Korea. (TEL) +82-42-869-3535 (FAX) +82-42-869-3510 (E-Mail) cha@cs.kaist.ac.kr (URL) http://dependable.kaist.ac.kr

Abstract

In this presentation, I will introduce why masquerade detection is an important security problem and how machine learning techniques, Support Vector Machine (SVM) in particular, are useful in developing effective mechanisms to detect masquerades. Experimental evaluation of the proposed approach on web server logs indicate that activities by masquerades can be detected while maintaining a reasonable rate of false alarms.

Short Biography

Sungdeok Cha: Sungdeok Cha is an associate professor in the Computer Science Division at KAIST and the director of KAIST IT Academy. His current research interests include software engineering and computer security. In particular, he has been working on formal methods in requirements engineering and intrusion detection research.

Before he joined the faculty of KAIS in 1994, he was a member of technical staff at the Aerospace Corporation in El Segundo, California and at the Hughes Aircraft Company in Fullerton, California. He received BS, MS, and Ph.D. degrees in computer science from University of California, Irvine in 1983, 1986, and 1991, respectively. He also taught at the California State University, Long Beach and Northridge, for five years.