Research On Intellectualized Screening System Of Network Information

The network information doped with much adverse information, not only affects the people's work, but, what's more, endangers the National security and even the sovereignty of our country. By the integration of advanced computer technology, this article analyses an intelligent network information screening system which can filter the adverse information and the sensitive information harming to the political, economic, military, cultural and other fields. On this basis, the papers provide some ideas and proposals in ensuring information security technically. In the system analysis, the www protocol is mainly studied.In overall architecture, original shape of screening system is realized, relevance technologies are studied:The host computer to be screened is defined as the screening client computer, and the client screening agent operates in the screening client. According to the constantly updated keyword list from the server terminal, the client screening agent collects, analyzes information in the information system. On the network sites, ip addresses, and generation time which consistent with the keyword information are encrypted and then submitted to the screening server terminal, which stores the received content and alarms on demand.The client screening agent which uses the intelligent agent theory for reference, composes spider, information storage, pre-processor, lexical analyzer and client agentupdate programme; the screening agent is mandatorily installed through legitimate means, and operates in the screening client terminal. The screening server consists of keyword server, screening result receiver, store, alarm and update center. The screening agent uses improved Spider screening technology to collect the information from the screening client terminal. Under the screening keywords, based on the lexical analysis principle of the seman -tic unit expressing tree pruning, the keyword Si=W_i,1W_i,2…W_i,n (where n is the length of the keyword Si) is transformed into the expansion formula of S^*i=W^*_i,1V1W^*_i,2V₂…W^*_i,n-1V_n-1W^*_i,n (where W^*i,k∈U(Wi,k) is the closure of transform Set, k=1,…,n;Vj∈V transform Set, j = 1, 2, ..., n-1), and then the past simple shielding keyword Si in the text is extended to any string of the shielding S^*i this model. According to the method that the semantic unit expressing library is transformed into the unit expressing tree, likewise, S*_i= W*_{i, 1}V₁W^*_{i, 2}V₂…W*₍i,n-1Vn-1W*i,n can be transformed into the expressing tree.Taking a sentence as the unit, take out the text messages to be filtered in order; for each sentence, take out each character in it; for each character, contrast whether it exists in the U transform library, if there is, the corresponding key character Wi, j will be obtained; take Wi, j real amount as the beginning keyword expressing tree and use the rapid pruning algorithm for pruning of all the keyword expressing trees taken out based on Wi, j. if there are some keyword expressing trees not to be pruned in the end, it means the sentence contains the keywords to be filtered.The LVS cluster system is chosen for storage of screening results. The LVS entire system can easily be expanded without resetting the whole system and without interrupting service. The expansion system is transparently operated by the terminal users, meeting the storage requirements of the intelligent network information screening system server. LVS can operate in kinetic energy in UNIX, BSD and SOLARIS systems.There are three types of LVS: VS/NAT, VS/TUN and VS/DR. At 100M, under the normal network service environment, it is assumed that the average data flow of each link is 10Kbytes, and the number of connections processed by VS/NAT per second is 1139.2Connections/Second. The maximum throughput of VS/DR or VS/TUN scheduler is 25,000 Connections/Second.This article solved three essential technical issues which include collection of information, lexical analysis, the result storage, realized the system prototype. The test result of the intelligent network information screening system shows that it can achieve to reduce manual labor, improve screening accuracy and efficiency.