Research On Malicious Code Detection Technology For E-mail System

In recent years, attacks in the network organization, seeks profit, professional, and orientation of the context of continuing to strengthen, viruses, worms, trojans and other malicious code on the network's communications activities have become more frequent, the technical means used by developing and evolving these allow normal network applications face significant security threats.The number and variety of malicious code is increasing, coupled with the rise of code obfuscation technology makes more and more difficult to detect malicious code. The traditional signature-based detection technology has been widely used commercial anti-virus software, but it must be in obtaining the signature of a class of virus to be effective after the detection of such viruses, and the signature after the infection usually is acquired. This feature allows the computer system the possibility of malicious code threats increase. Although there have been a variety of malicious code detection method, with mixed success, especially for detection of unknown malicious code is still a hot issue to be resolved. In this paper, are already testing a variety of malicious code detection algorithm is carried out in-depth research and analysis, and proposed standards for sparse coding based on malicious code detection.The concept of sparse coding comes from the study of visual neural network , it is a neural network method for finding a representation of multidimensional data in which each of the components of the representation is only rarely significantly active.Sparse code theory establishes a scientific quantitative link between the information processing mechanisms of visual neurons and the statistics of input visual stimuli,and provides an efficient tool to understand the neurainformation processing mechanisms.It has been applied in blind source separation , speech signal separation , image feature extraction , naturalimage denoising and pattern recognization , and it has achieved many results and has important practical value.This article will first apply sparse coding areas of malicious code detection, using the sparsity of feature vectors, the structural features of malicious code analysis and classification, achieved good results, not only can detect known malicious code detection, on unknown malicious code detection has some ability.