Research On Technology Of Network Information Monitoring And Blockading

With the coming of information era, the mass information on the Internet bringsmuch convenience to our work, study and life, but there is still much sensitiveinformation on the Internet which contains erotic, retroactive and violent contents,and brings negative influence to the youngsters even to social stabilization. Thetechnology of Network Information Monitoring and Blockading finds out illegal textthat contains sensitive information and to block it.To create a healthy network environment, this article studies the principle ofnetwork data packets capture and analysis, information filtering, and applies thesetechnologies to the blockade of sensitive information. Network data packets can becaptured by means of WinPcap on the platform of Windows, and be analyzedaccording to the TCP/IP protocol to recover the Web text.In order to speed up the system and to increase the efficiency of the blockading,this system uses two level filtering policies. Namely, the 1st level is based on the IPaddress and key word, and the 2nd level is based on the filter policy of the textclassification. Legal IP address database is used to decrease system's misplay.After the brief introduction of research background as well as domestic andinternational, the Vector Space Model is used to express text, and K-Nearest Neighboralgorithm (KNN) is used to identify sensitive information.Use the Forward Maximum Matching to set up the Chinese text segmentsdatabase according to the Web text in the text set. Summarize the feature of illegaltext and express the text by the Vector Space Model. At last the K-Nearest Neighboralgorithm is used to determine the tested Web's classification. The universalestimation method of filtering performance is used to estimate the experiment data,the result shows that this system's performance is satisfying.If the Web text contains sensitive information, the TCP connection will be cut offby sending a FIN message to block up the spread of sensitive information.