Research On The Techniques Of Network Information Security Audit Based On Content

With the rapid growth of Internet, our life is tied to Internet more and more tightly. Governments, companies, and even individuals rely on Internet to communicate with others. Internet is now an indispensable element of our everyday life. However, people have realized that Internet is not as good as our thought when they enjoy the convenience. Not only news reports, technology dynamics, public entertainments and other useful information are on the Internet, but there is a variety of illegal information, such as erotic, violent and evil things, on it as well. With such information, people have been disturbed very such in many ways. Moreover, these things can be a huge trouble for the country and society. How to audit those wicked information so that they could be a problem to us? As an important component of the Internet securing, securing content of the Internet information has been attached more and more importance these days.Security Audit is a very attractive approach with real-time, dynamic and active defensive characteristics. In the relevant international standards of information security, network security audit has become an indispensable component part in constructing safety network. But the auditing products that are strict accordance with the definition of information security are very little.Content-Based Network Security Auditing System (CBNSAS) would be a specialized system, which analyze and audit the network's information content. At present, the research on CBNSAS is a hotspot in the field of network security; lots of key technologies are still in their childhood at home and abroad. This thesis researches on the technologies of content-based network security audit, research work is carried out in two major areas.1. With the view of key technology for network information content securing implementation. The key technology mainly includes: Chinese segmentation, text classification, intelligent multimedia information recognition, and data storage technology. Text classification technology is an important component of content security, and now also a hot issue in content security research. The thesis mainly focuses on text content research, gives a deep analysis of text classification technology, and so improves text classification algorithm—Support Vector Machine. As to one-class SVM, coupled with genetic algorithms, the thesis proposes one-class SVM based on GA which be applied to information audit. 2. With the view of security audit system design. Under the circumstances of sufficient network bandwidth, the processing speed of content analysis module will become bottlenecks of system performance. Furthermore, the processing speed and the accurate identification rate of sensitive processing speed and the accurate identification rate of sensitive information is a contradiction, which cannot be both taken care of. On conducting network security auditing system, double layers structure is used to improve the speed and accuracy of the audit as possible.In addition, the thesis gives an overall analysis of newer network security tchnologies for network security information audit, such as introduction of concepts, performances and comparing with other tchnologies.Finally, research and design of content-based network information security audit system. And the thesis introduces the various modules of the system in detail.