A Study On The Application Of Personalized Information Retrieval System Based On Ontology

With the growth of Internet, the on-line resources are developed in the index speed,and simultaneously people's need for information is increasing imaginably. However, it is difficult for people to obtain what they need from the information sea in a quick and convenient way. As a result,the problem so-called"information overload"and"information confuse"appears. The most important reason is that the personal factors for each people are not considered by the general search engine, and provide the same results. Therefore, how to learn about the user intention,obtain the personalized information and provide personalized service for the user have become the focus in the field of information retrieve.The dissertation constructs a Personalized Information Retrieval System (PIRS). It integrates traditional information retrieval systems and user model and collection, so as to establish new type of user interest orientated information service system. The PIRS can obtain the identification of user interests,the learn of user model automatically, query expansion and measure the match between returned results and user interests based on the user model,consequently providing personalized service. The following are the main contents from this research:Firstly,the dissertation discusses the present situation of the information retrieval and analyses existent problems therein,then the research significances are proposed. According to what is mentioned above,the author discusses the procedure of personalized information retrieval and makes the frame construction and overall design of PIRS.Secondly,based on the stage of constructing the user model,the article researches including: (1) The construction of domain ontology. The related words are obtained by text mining, and then domain experts construct the domain ontology semi-automatically. The construction is implemented with the tool of text mining and ontology constructing. (2) The construction of initial personal ontology. Based on the information submitted by user,the initial personal ontology is constructed by ontology projecting and its algorithm description is listed. (3) The update of user model based ontology. The interest of users will be constantly changed by time decease. Therefore,we need to construct reference ontology by analyzing the browsing or retrieval process, and then merge it into personal ontology to complete the personal ontology learning,which is used to update the user model .Thirdly,user model is applied into the personalized information retrieval system,which contains two aspects: (1) Query expansion. Base on the three kinds of query,expansion algorithms are designed correspondingly to make query term reflect the needs of users well. (2) The filter and optimization of query result set. The correlation of document and query term is calculated by Cosine theorem,then documents whose correlation is large than the assumed threshold are presented to user in descending order.Fourthly,the article designs a experiment to prove the feasibility and validity of PIRS,in which the search logs are provided from the Sogou laboratory. The experiment result indicates that compared with traditional information retrieval system,the recall and precision ratio are improved by PIRS.