Research On A Pattern Matching Algorithm Based On Word Frequency

The more rapid and efficient of technology such as firewall, VPN, PKI and intrusion detection is required by the increasing complexity of Internet environment and the increasing number of Internet. Pattern matching which is able to effectively support network content security and improve the performance of network equipments is one of the most important high-speed network.In this thesis, the research background, development and current research status of pattern matching are written first, followed by the related technology of content security as well as firewall and intrusion detection. After that, typical pattern matching algorithms are described and analyzed. To improve the shortage of pattern matching algorithms, a pattern matching algorithm based on word frequency, which is named BCFM, is proposed here. This algorithm establishes a character frequency table firstly, then finds the lowest frequency character and the second lowest frequency character according to this table, and records their relative position. When the pattern matches, it will be find out the lowest frequency characters in the text firstly, then the character on the relative position with the second lowest frequency character to match directly. The matching process is completed quickly. BCFM algorithm is more accurate for character positioning than other algorithms, so it increases the efficiency of pattern matching.Firewall and intrusion detection is also investigated here. Finally, an experiment is completed to test and compare the performance of BM and BCFM. The results indicate that BCFM is provided with preferable efficiency on time. It is more efficiency when pattern is shorter and text is longer.