Research On Improvement Of Search Algorithm Based On Web Similarity

Along with the rapid development of the internet, the amount of information in the network increases exponentially, it makes the user to access information becomes more and more difficult. In order to heterogeneous resources better use of the Internet, search engines emerge as the times require. Usually the standard of evaluate search engine performance is search engine users satisfaction, while the user is using the search engine to search, generally preferred click ranking on the relative front webpage, so the search engine results for a reasonable sort will significantly enhance the user experience of search engines. Page Rank algorithm is widely used to measure the importance of web pages, but the traditional Page Rank algorithm ignores some factors may influence the importance of web pages in the calculation process, there are many defects.This paper studies the similarity of web-based Google’s famous Page Rank sorting algorithms. Firstly, this paper expounds the present situation of research on Page Rank algorithm and Page Rank algorithm of the background and significance of research at home and abroad, introduces the development course, search engines work and the judgement standard, then analyzes the principle of Page Rank algorithm. The classic webpage link analysis algorithm Page Rank “each link represents a webpage author an independent accreditation” pointed to the webpage as a precondition for the algorithm, but one of the main defects of the traditional Page Rank algorithm is a webpage Page Rank average weights assigned to all of the chain, and not consider the semantic information of the webpage. We put forward an improved Page Rank algorithm based on the similarity of the page, through the similarity weight Page Rank weight distribution, similarity contains two parts of text similarity and similarity of web page links. With consideration of the chain and the target web page similarity information, which not only increased the importance of web pages of the accuracy, precision and the results retrieved is higher.Finally, in order to verify the algorithm performance and efficiency. In the part of experiment, with the help of the open source search engine Iveely please some users experiment with the real test in the Internet environment. Small-scale user testing results show that: the improved algorithm into webpage text similarity and webpage link similarity, improve the precision of search results and user satisfaction.