Download presentation
Presentation is loading. Please wait.
Published byGeoffrey Scott Modified over 8 years ago
1
Intelligent Database Systems Lab Presenter: CHANG, SHIH-JIE Authors: Bui Quang Hung, Masanori Otsubo, Yoshinori Hijikata, Shogo Nishida 2010.WIA. HITS algorithm improvement using semantic text portion
2
Intelligent Database Systems Lab Outlines Motivation Objectives Methodology Experiments Conclusions Comments
3
Motivation Previous researches have tried to solve following problems using anchor-related text. Link-spamming problem BHITS method Automatically generated links, banner ads => Topic drift problem Identify important link => Chakrabarti’s method P P P P P Page A Authority score A P P P P P Page B Hub score B
4
Intelligent Database Systems Lab Objectives Investigate the effectiveness of using Semantic Text Portion (STP) for improving the HITS.
5
Methodology – The HITS algorithm authority hub Root set R Base set i
6
Methodology – The BHITS method authority hub Root set R Base set i hub_wt auth_wt
7
Methodology – Chakrabarti’s method authority hub Iteratively calculates authority scores and hub scores.
8
Intelligent Database Systems Lab Methodology – Chakrabarti’s method
9
Intelligent Database Systems Lab Methodology – Semantic text portion(STP) STP is a text portion in the original page which is semantically related to the anchor pointing to the target page. LSP: Local Semantic Portion USP: Upper-level Semantic Portion
10
Intelligent Database Systems Lab Methodology – Example of LSP 410list
11
Intelligent Database Systems Lab Methodology – Example of USP USP
12
Intelligent Database Systems Lab Methodology-
13
Methodology – Collecting base set I 1 Root set R Base set i
14
Intelligent Database Systems Lab Experiments
15
Intelligent Database Systems Lab Experiments
16
16
17
17 Ranking results for the architecture query
18
Intelligent Database Systems Lab Ranking results for the bicycling query
19
Intelligent Database Systems Lab
21
Conclusions The use of STPs is best for improving the HITS algorithm.
22
Intelligent Database Systems Lab Comments Advantages - Effective. Applications - Web mining 、 Rank web pages.
Similar presentations
© 2024 SlidePlayer.com Inc.
All rights reserved.