Study On Algorithms Of The Vertical Search

With more and more information appears on the internet, the traditional general search engines for information search gradually revealed more and more problems, like"accuracy lackness, poor timeliness,low coverage". The General search engine can satisfy the user's public infomiation query demands,and it according to user input index page query strings in content with the matching degree level,and return to the page. This kind of methods not only has brought lack of accuracy and low page effectiveness, and already cannot satisfy people growing's individualized service needs.In view of the above situation, another kind of search engine, it can provide more satisfactoried results than traditional search engine in certain scope,which is called vertical search engine. The core of vertical search is vertical search algorithm.First, this paper introduces the basic principles of universal search algorithms and the key technologys, and then analyzes the advantages and disadvantages of them,and proposed in this foundation the new vertical search algorithm. This paper uses PageRank algorithm and the Hits algorithm to improve:increase the proper weights of subject keywords and give related properties proper proportion weights. In traditional search algorithms produce "topic drift" and "attributes drift" roblems,the paper adjuste the weighting factor in the algorithms, as far as possible to avoid this kind of problem. In this paper algorithm threshold is discussed, and points out the thinking actors and the factors affect the results of its algorithm.Based on the improved algorithm we have designed a simple experiment environment, and the results of the improved algorithm was validated, and the result shows that,the improved algorithm can partly avoid problems such as topic drift.