Research On Search Engine Ranking Algorithm Based On Link Analysis

With the rapid development of the Internet, Web has become an important way to obtain information for people. However, with the number of web pages increasing, the explosive growth in information, brought inconvenience for people to find information. How to retrieve the really interesting information rapidly and accurately is the target that the major search engine companies constantly improve technology. The introdutction of Web data mining improved the retrieval quality. Especially the applications of Web structure mining in searching engine helps user quickly find the useful information from the search results very well. HITS algorithm and PageRank algorithm, which are the classical algorithms on Web structure mining, are researched in this article.According to PageRank algorithm we study its basic thought and discuss the influence on the links structure to the PageRank value of the web page, and then propose corresponding strategies to raise the rank of the website. Based on the analysis about the disadvantages of PageRank algorithm, an improved algorithm based on the correlation between query topics is proposed, and experimental data indicates that the algorithm is effective. For HITS algorithm, another sorting algorithm in search engin, those are studied that its idea, implementation steps and disadvantages. Based on the analysis about the main reason of topic drift, an improved algorithm based on the relevance between pages is proposed.Through a group of comparative experiments, the superiority of the improved algorithm is verified.