An Improved Method To Rank The Web Pages Based On Temporal Information

The information has made it's remarkable progress in recent year,the dependent from users on the information warehouse is also growing sharply,especially on the search engine. And on the other hand,the internet can provide more and more kind of knowledge to the users. In such a huge and explosively developing information world,finding and ranking a tiny web pages from abundant pages becomes the fetal features of information retrieval tools.There are now mainly four kinds of analysis method,one is to analysis the linkage structure,the second is to analysis the content,the third is to use user log,and the last one is to combine the above three.As the most different features web pages have is link structure,so the first one is most important to information retrieval on the internet.However most of the currently used link analysis methods ignore the temporal information's importance.That is to say they are using a 3-D way to illustrate the web and do not use the 4^th dimension which is time.This problem has been first discovered by Albert-L(?)szl(?) Barabasi,and be implemented by Philip S.Yu and developed several times by the other researchers.But they all based on the fundamental of summation of links in different time period,without considering the key features of web pages.After all,this thesis has done works as follows:This paper introduced several algorithms using linkage,implemented web crawler system by java,improved the former Philip S.Yu's algorithm,introduced a novel method using the changing feature of web and analyzed them mathematically,set up experiment platform using MATLAB,and compared the result created by them.The method analyzing the change of web pages is firstly proposed.Compared with the former algorithms,this algorithm can more accurately describe the web by temporal dimension,and changed from simply sum the exponential result of each link to more deeply analyzing the web's state changed between time periods.This novel method has improved previous method significantly.