Research And Implementation Of Page Rank Algorithms On Search Engine

With the rapid development of the Internet, the documents or web pages on the Internet grow very fast. According to some related information, Internet resources increase more than doubled every year. In this backdrop the search engine come into being which helps users to access information from a lot of information resource library. Therefore, it is an important means for people to collect information resources by using the search engine. While relying on the search engine's wealth of information resources, users also put forward higher requirements on the quality of service provided by the search engine. The result of the search engine directly affects its efficiency and popularity. One of the core technologies is how to sort the results pages.The technology of the search engine has developed rapidly in recent years. The most representative company is Google, and its rapid development has also stirred up other colleague's enthusiasm for their study. But the search engine returns so many query results sometimes that users can't read all the summaries. Some statistics show that the first three pages of the results list are most frequently browsed by user. Of course, the first page's frequency is the highest, and there is a close relationship between the search engine's quality and its results'reasonability. So people's study on the page rank algorithm is gradually in-depth.At the first, this paper introduces the working principle of search engines, as well as the development process, and then mainly makes a study on the ranking algorithm of some well-known search engines, analyzes the impact factor of the results sort. The paper compares the advantages and disadvantages of several sorting algorithms and the algorithm's suitable situation. This study also focuses on the PageRank algorithm, Hits algorithm and Frequency position weighing algorithm. Using the cosine value of the vector angle between document keywords and query words, the existing vector model is improved in this paper. Although the improved model needs more computation, the similarity between the document and query words can be more precisely calculated.This paper analyzes the reasons for the shortcomings of Frequency position weighing algorithm. Then on the core idea of the algorithm the paper designs a formula for calculating the document weight value by combining with the inverse text frequency index and the improved vector model. The process of the formula's calculation is effectively integrated with the query words. The system implementation and evaluation proves that this algorithm is easy to accomplish and to sort the results and is more accurately and efficiently.