The techniques for computer security intrusion detection based on Preserving Embedding for Anomaly Detection

Abstract

Computer security has been attracting more and more attention ,since intrusion detection have become a significant threat in recent years. The techniques for intrusion detection are generally classified into two categories, which are anomaly detection and misuse detection respectively. In this paper, we mainly focus on anomaly detection on behavior of process which is in the form of system call traces. Each process trace is recorded by different system calls that can be naturally deemed as high dimensional data, as one operating system may have a great deal of different system calls. Thus, it is natural without any doubt to say that dimension reduction technique has the opportunity to make a better performance improvement by exploiting classifier in the low dimensional subspace.