| Because of the continuous development of Internet technology, more andmore new websites with feature-rich occur, the reason that people surf Internet isnot only to read the news, or to search information, but also to record themselvesin the network’s daily life. With a brief state to express their feelings, or views onsomething. Network is not just to present data to the platform, and become awindow to for users to show themselves.This data created by the users is different from by professional editor.Because it is more free, flexible, richer and rapid. So the research on this data isfull of value. However, the characters of the data bring some new problems forthe traditional search engine.This paper first divides search engine into for parts, crawler, indexing, queryprocessing and data display. Then it analyses the problems in getting the data foreach module. We put forward a new theory and solutions for these difficulties.In the web-crawler section, academic community that the web page changesfollow a Poisson process, but this paper analyzes the web pages’ changes indifferent time, and add mutual intimacy between users to law of the change, thenproposed a new web page changing model. In indexing section, we use a varietyof index that can not only improve the timeliness of the results, but also cansupport for the period of the statistical data query. In data sorting section, weimprove the page rank method by taking the new characters like comment andsharing in social network service. In data showing, we use emotions to classifythe results of data in order to display to the user a more intuitive data.After that we design and implement a new sns search engine based on basisof these solutions, the experimental results verify that the system has goodperformance at the end of the paper. |
