topic distillation;
HITS;
hubs and authorities;
diffusion of authority;
the TKC effect;
D O I:
暂无
中图分类号:
TP18 [人工智能理论];
学科分类号:
081104 ;
0812 ;
0835 ;
1405 ;
摘要:
This paper presents a variation of the HITS algorithm in an attempt to overcome some of the limitations of the original algorithm The proposed algorithm employs an augmented adjacency matrix of the link structure in which some virtual links are added to the original graph in order to connect directly each node to nodes only two links away from it. The intuition behind our approach stems from a personal belief that the references used in an authoritative page should be treated as even more authoritative than the page itself We compare the results of applying our modified HITS to the results derived through HITS on broad-topic Web queries (both single-topic and multiple-topic). These comparisons show that the proposed algorithm reduces the effects of two drawbacks of the original algorithm, mainly the diffusion of authority across siblings and the well-known Tightly-Knit Community (TKC) effect problems.