Research On Technology Of Topic-Oriented Information Searching And Filtering

With the rapid development of information on the Internet, around 20,000,000 web pages appeared, and lots of them have little entropy of intelligence. General search becomes weaker and weaker. The precision ratio and recall ratio have been declining that some special users have to look for a new method to fit their needs. It has become a tendency to develop specific Internet information retrieval tools.The technique of topic-focused web information searching and filtering described in this paper. According to the personality of topic-focused information, we present a new Topic-focused method including information query representation, information retrieval and information filtering. The core technique is how to represent the query of information, how to search the relevant information and how to filter useful information.Referring to national 863 high-tech Project, the main content in this paper was introduce as follows:1. Representing personal information query, which make user no more fretful about useful keywords. Users can provide keywords more easily, customize specific topic and retrieval vector, which match the information user is looking for.2. Introduced Topic-based query topology, which broken out the quantitative limitation of keywords and"word-mismatch". The method make the loss of the relevant information decrease and make the recall ratio go up.3. Controlling of the Anchor-depth can make the system work more efficiently not search anchor greedily. It make the system work in high crawl efficiency4. Improving on combining content-based and anchor-weight-based filter algorithm, which resolve the problem of"topic-drift". It works well on popular PC and achieves high topic-focused information recall ratio.