Research On Page Value Computing Algorithm Based On Content And Web Link

This Dissertation describes the model for setting up the personalized informationof users under the Internet environment. It draws the methods of optimizing the model.The algorithms of computing the value of pages are considered. It designs the subset ofdata for testing, which was part of the WT10g. The subset keeps on the characters ofInternet. The subset of data is used to test the algorithm, and the results show thealgorithms are suitable for the model in this dissertation.This dissertation introduces the methods of computing the value of pages by usingthe content and the links of the pages. According to that, a set of pages isrecommended to users. It helps the user to get the better precision and recall.According to the number of pages are considered in the computing, the algorithm isdivided into two classes, the one is based on Content and Link Based Fast RankingAlgorithm, and the other is based on Content and Link Based Complete RankingAlgorithm. The results of simulation show that even part of pages is considered incomputing, the recommend sets of them are similar.The algorithm, which was drawn in this paper, could be used in the specialdomains and some IR systems. It solved the difficulties, such as huge volume,heterogeneous, dynamic and semi-structured in IR process. It supported the user to findthe value pages quickly.The last part of this paper points out the rooms of improvement and the futureworks.