TY - CHAP
T1 - Exploiting pagerank at different block level
AU - Jiang, Xue-Mei
AU - Xue, Gui-Rong
AU - Song, Wen-Guan
AU - Zeng, Hua-Jun
AU - Chen, Zheng
AU - Wei-Ying, Ma.
N1 - Publication details (e.g. title, author(s), publication statuses and dates) are captured on an “AS IS” and “AS AVAILABLE” basis at the time of record harvesting from the data source. Suggestions for further amendments or supplementary information can be sent to [email protected].
PY - 2004
Y1 - 2004
N2 - In recent years, information retrieval methods focusing on the link analysis have been developed; The PageRank and HITS arc two typical ones According to the hierarchical organization of Web pages, we could partition the Web graph into blocks at different level, such as page level, directory level, host level and domain level. On the basis of block, we could analyze the different hyperlinks among pages. Several approaches proposed that the intrahyperlink in a host maybe less useful in computing the PageRank. However, there are no reports on how concretely the intra- or inter-hyperlink affects the PageRank. Furthermore, based on different block level, inter-hyperlink and intra-hyperlink can be two relative concepts. Thus which level should be optimal, to distinguish the intra- or inter-hyperlink? And how the ratio set between the intra-hyperlink and inter-hyperlink could ultimately improve performance of the PageRank algorithm? In this paper, we analyze the link distribution at the different block level and evaluate the importance of the intra- and interhyperlink to PageRank on the TREC Web Track data set. Experiment shows that, if we set the block at host level and the ratio of the weight between the intra-hyperlink and inter-hyperlink is 1:4, the retrieval could achieve the best performance. © Springer-Verlag 2004.
AB - In recent years, information retrieval methods focusing on the link analysis have been developed; The PageRank and HITS arc two typical ones According to the hierarchical organization of Web pages, we could partition the Web graph into blocks at different level, such as page level, directory level, host level and domain level. On the basis of block, we could analyze the different hyperlinks among pages. Several approaches proposed that the intrahyperlink in a host maybe less useful in computing the PageRank. However, there are no reports on how concretely the intra- or inter-hyperlink affects the PageRank. Furthermore, based on different block level, inter-hyperlink and intra-hyperlink can be two relative concepts. Thus which level should be optimal, to distinguish the intra- or inter-hyperlink? And how the ratio set between the intra-hyperlink and inter-hyperlink could ultimately improve performance of the PageRank algorithm? In this paper, we analyze the link distribution at the different block level and evaluate the importance of the intra- and interhyperlink to PageRank on the TREC Web Track data set. Experiment shows that, if we set the block at host level and the ratio of the weight between the intra-hyperlink and inter-hyperlink is 1:4, the retrieval could achieve the best performance. © Springer-Verlag 2004.
UR - https://www.scopus.com/pages/publications/33745882026
UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-33745882026&origin=recordpage
U2 - 10.1007/978-3-540-30480-7_26
DO - 10.1007/978-3-540-30480-7_26
M3 - RGC 12 - Chapter in an edited book (Author)
SN - 3540238948
SN - 9783540238942
VL - 3306
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 241
EP - 252
BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PB - Springer Verlag
ER -