| In recent years, the rapid development of the internet brings convenience to experience network, but unprecedented challenges to finding useful resources as well. Currently, people search resources on the network mainly rely on search engines, such as Google and Baidu, which are powerful and usually pleased by the users under normal circumstances. But when users want to query information about a particular topic, this type of engine is somewhat limited because they are based on the linked Web page sort algorithm and do not fully take into account the relevance of Web page content.Typical algorithms based on link analysis is PageRank Algorithm. This article analyses the correlation of page content passed when the PageRank index is passed, it first expounded the development and current status of search engine, introduces the concepte and classification of information retrieval model, analyses several sorting algorithm and describes their principle and evaluation, and then explaines the traditional PageRank algorithm and several current improved algorithm. Based on the above analyses, it raises a new algorithm, which does not change the traditional algorithms for Computing Web page rankings, but also gets traditional PageRank algorithm for the calculation of PageRank value by smaller d index.The new algorithm MPR provides broaded space for extending PageRank, and by choosing the appropriate parameters experimentally proved to be capable of achieving better page ranking results than traditional PageRank Algorithms. |
