Hyperlink classification: A new approach to improve PageRank

Hyperlink Structure is widely used in the hypertext classification, but it has not been paid enough attention. We propose a hyperlink classification approach to improve PageRank algorithm which is widely used in the link analysis of search engine. The cause of the topic drift problem is analyzed and the hyperlinks are classified according to their creating motivations and effects. The improved PageRank algorithm is implemented on the open source search engine NUTCH in Chinese Internet. The experimental results show that the improved PageRank algorithm performs better than the standard PageRank.