Research And Design Of The Individualized Network Document Retrieval System Based On User's Model

This paper has put forward the design research of a kind of individualized network information of document retrieval system,through collecting users' basic information and interest information,set up the new-type individualized network information service system facing user's interest.It can be said on the individualized foundation that realizes of the system in users' interest model,search engine,the interface of system and search engine,and vectorial match of model algorithm system individualized base that realize.This paper has also done the introduction of studying separately on discernment of Web page,cluster's technology.In the article except that have built a user's model according to certain method and principle,also act as relevant research on advanced Agent and machine of intelligence technology in modeling at present.This paper finally introduce systematic another important function also --Push voluntarily serving in information retrieval.The realization that system's serve of individualized network document information retrieval can be go on by five steps following mainly:The first step,users input one's own keyword or phrase information in the search frame of the system user's interface,then turn towards and refer to the system.Second,send the keywords that obtain from first step to the special search engine through search engine interface of system,they carry out the real search behavior.Gain the document's characteristic through collecting algorithms at the same time. The third step,calculate out relevant degree the document vector quantity of every result obtained through the vectorial model algorithm and user's model vector quantity,and the way to arrange in the order of relevant degree,then save the vector quantity of the document. The fourth step,the result shows.In user interface,accord with user result of model,show according to maximum and relevant degree that user establish by high finally order.The fifth step,according to users' feedback behavior,optimize user's model further,adjust the relevant right value of one and right value factor.For this paper and systematic realization of goal,the thesis organize discussion to study with seven chapter:Chapter one is introduction part,start with network information as we obtain main way of information now,this chapter discuss the main way of we obtain the information of network,and the problem faced while obtaining the information of network at present,background drawing the thesis' studying.And describe the definition,background,and current situation of individualized network information retrieval serve at present.The most direct concept of the individualized network information retrieval serve,is with the service mode that " users are as the center ".Already according to clear request that user put forward,interest model of user search through user individual character, analysis of habit in the network, offer information and service that may need to users, train the individual character,guide the service of the demand. Its object is to respect users,study users' behavior and habit,choose the even more important resource for users,offer the service of the characteristic network information retrieval.Chapter two is the study on individualized network document retrieval technique.First of all,this chapter introduce some individualized concept, thought and individualized in this paper characteristic of the research involved in the individualized technology.Individualization is to make thing have distinction,enable its distinction show especially.The individualized thought is to regard systematic user as the center,in the meaning originally in the paper, refine its interest and partiality,based on this and present the information searched out from the network in front of users passively and voluntarily.The essential feature in the system of individualization has the highest user's needs, users describe the file (user profile ) ,regard existing search engine as tool,based on re...