Font Size: a A A

Study On Web Information Credibility Evaluation Method Based On Improved PageRank

Posted on:2012-04-19Degree:MasterType:Thesis
Country:ChinaCandidate:W Y MaFull Text:PDF
GTID:2178330338995361Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
In recent years, with the rapid development of Internet, Web provides users with a large number of information resources, and gradually has become an important way to obtain information. However, it is increasingly difficult for network users to screen the web information with higher credibility because of the web information's rapid increase.This thesis mainly studies the academic problem about how to evaluate the credibility of the web information based on the traditional web structure mining algorithm PageRank.Through practical observation and detailed analysis, we find that there are some factors which can influence the credibility of the web information. Such as, the interactive structure of web pages, the correlation degree between two web information themes, the release time of the web information, and so on. This thesis takes into account the above factors, and put forward a web information credibility evaluation method based on improved PageRank. The conceptualization of this method is that: first, consider the interactive structure of web pages, and create the interactive chart of web information by analyzing the web page number of linking and being linked to the web pages which contain these web information; then, express the web information theme as the form of a vector by the TF-IDF formula, calculate the correlation degree between two web information themes using the vector cosine distance formula, and analysis the relationships between two web information themes; at last, introduce the time decay function into the process of the web information credibility evaluation according to the different release time of web information, and reflect the effect degree of the release time to the credibility of web information using this function.In this method, it is using the interactive chart of web information to compute the credibility of the web information. The calculation method is to introduce all relevant factors into the interactive chart of web information, and maintain the credibility of the node by the trust propagation mechanism to achieve the purpose that the credibility of other relevant nodes dynamically change with the change of one node's credibility.This thesis designs the experiment to validate the performance of the evaluation method in this thesis, and the experiment results show that this evaluation method can provide users with more credible and valuable web information.
Keywords/Search Tags:PageRank, Web information credibility, Relation-degree, Time-degree, NPR(New-PR)
PDF Full Text Request
Related items