Intelligent Search Technology Of Network Information Based On Military Application

Information gathering, processing and analysis relate greatly to all aspects ofnational development and advancement. In addition to factors such as the strategyand tactics, technical equipment, the weaponry, and education level, the capabilityof collecting information is a very important one for the evaluation of the battleeffectiveness of a nation's military forces. With the emergence of theinformation based "third wave war" time, high tech trend and outstanding abilityto collect infomation have become a major direction for the modernization ofnational defense. With nations'competition becoming increasingly fierce in allfields, information gathering and analysis have drawn more and more attention.Information collecting is not only the base and precondition of theinformation research, but also the substance base of information analysis.Collecting the public information in military domain is an indispensable part ofmilitary information collecting. With the rapid development of computer andnetworks technology, collecting the public information has become an importantway of information collection, on which is paid more attention by the information departments of every nation. However, the feature of Internet is free,heterogeneous, which makes the collection of valuable information with greaterdifficulty. In order to gather the valuable information, one must appeal to somesoftware tool. With the use of web search engine technology, this problem isresolved to some extent, but many new problems come forth such as "informationoverload", "poor correlativity", and so on.Our work is to develop a network based information search system onmilitary application using meta search, information extracting and noiseseliminating, Chinese words segmentation and ambiguity removed,vector spacemodel and other technologies. This system can automatically searche and analyzethe public information from Internet based on the key words defined by user.Intelligence information will be automatically analyzed. The information whichhas high relation with the request of user will be collected in the form of link andstored into information databases.The main works is:1 . Search technology.Based on the analysis of search engine structural framing, a networkinformation collection system (Network Information Search Finder, NISF) isproposed and developed. The system works in the condition of uninterruptedmovement, fixed time searches and collects the public information on Internet.The NISF system including the user interface, the standard search engineassignment and the transfer mechanism returns the search result which areprocessed to user.In the design of user interface, user demand model is constructed and aninformation user model is proposed based on keyword list and user's feedback. The system can adjust the weights of keywords by continuously collecting theinformation of feedback from user. Threefore the varacity of the information withthe request of user from the Internet is improved.In the mechanism of dispatch and call standard search engine, the conceptand mathematical model of dispatch coefficient is proposed. In view of users'search subject description, this mechanism can judge the search performance onthis kind of standard search engine by the dispatch coefficient, then choose themost superior standard search engin to complete the search task.In aspect of the processing and return the search result, the databasetechnology is used to eliminate the informations which havs same title or sameURL in the search result, and the VSM is also used to analysis and computedegree of correlation between search result and the user demand model. Animproved self adapted text filting algorithm based on the user feedback is putforword. In aspect of the extracting the reach information, the subject links basedon the HTML mark and key words, the algorithm of the subject text contentextraction and evaluation method as well as the mathematical model of the resultsof the Web page text content extraction are proposed.2. Information processing technology.Our task on information processing is to realize the automatic classificationof the information dobucments. At present, because computer can't understand thenatural language completely, the common method of extracting the highfrequency words from the documents is the one which is used to describe thedocuments. In the extraction of high frequency words, a Chinese participlemethod is proposed on the dictionary and the word frequency Chinese. Bycomparison vector space model with the set theoretic model foundation, vector space model is used to compute the similarity of documents. By comparison onthe performance of the cosine formula with that of the Euclidean space distance,the cosine formula is used to compute the distance between each document. Theclassification algorithm which is combined VSM, KNN and SVM unify isadapted to classify the documents. Then the results of classification are store inthe database.3. Development and realization of the software system.Based on above technology, the network information collection system (NISF)based on military application has been developed using Borland Delphi7.0. Thesystem runs under Microsoft Windows XP operating system.Conclusion:1. User requirement model which uses keywords and user feedback canquantify the command of users and reflect the search request of users vertbaly.2. Dispatch coefficient can primely estimate the capability of standard searchengine while searching a certain keywords. It provides a foundation formeta search system when calling standard search engine.3. The improved self adapted text filtering algorithm can give the highercorrelation information to users.4. The extraction technology of Web page information has good performancein extracting Web links and text content. Theme information rate serves as afoundation in estimating the extraction algorithm of Web page information.5. Chinese word segmentation based on dictionary and word frequency has anoutstanding performance in extracting high frequency word from document.Network information search finder based on military application isconstructed. It has higher martial features and satisfies the command of military forces. Data mining technology including automatic classification, informationfiltering, information extracting and automatic word segmentation makes theinformation processing more intelligent. At the same time, this system can searchboth martial information and other professional intelligence information. So, thissystem has not only the extensive application value, but also the significance bothin military and in realistic use.