SSTAG: A Novel Tag Recommendation Mechanism to Web Page Search

다국어 초록 (Multilingual Abstract)

As a social bookmark tool, Folksonomy gives high freedom to users and allows users to share and tag resources, such as web pages. However, many tags applied arbitrarily by users cannot really reflect contents of web pages and lead to ineffectiveness in information retrieval. Moreover, there are still some important tasks about how to eliminate ambiguity more easily and recommend more interested web pages to users. To resolve the above problems, we propose a novel mechanism, SSTAG, which can recommend a set of Super-tags to users for their choices based on the input keyword. As various topics related to the keyword, the Super-tags are selected from different clusters of web pages. A user chooses a Super-tag, which means the user may have chosen an interested category, and then some more detailed tags in the category will be recommended as Sub-tags. The relationship between Super-tag and Sub-tag is just like navigation and positioning. Likewise, the user can choose one Sub-tag and submit it with the Super-tag. By means of the user’s choice, this system can capture user’s preference and recommend a series of related web pages. We employ a real world dataset to examine the mechanism, and the experimental results show that this mechanism can eliminate ambiguity efficiently and recommend a set of high-quality tags.

목차 (Table of Contents)

• Chapter 1. Introduction 1
• 1.1 Background 1
• 1.2 Related Work 2
• Chapter 2. Processing Flow and Architecture 5
• 2.1 Processing Flow 5
• 2.2 Architecture 6
• Chapter 3. Tag Recommendation Process 8
• 3.1 The Selection of Candidate Web Pages 8
• 3.2 Tag Refining Module 9
• 3.2.1 Motivation 9
• 3.2.2 Tag Refining Method 11
• 3.2.3 Stop-words 12
• 3.3 Web Page Clustering Module 13
• 3.3.1 Similarity of web pages 13
• 3.3.2 Clustering Algorithm 17
• 3.4 Tag Ranking Module 18
• Chapter 4. Experimental Evaluation 21
• 4.1 The Selection of Dataset 21
• 4.2 Recommendation Results Analysis 22
• 4.2.1 Super-tags Recommendation Analysis 22
• 4.2.2 Sub-tags Recommendation Analysis 24
• 4.3 Ambiguity Analysis 25
• 4.4 Web Page Search Results 27
• 4.5 Performance Evaluation 29
• Chapter 5. Conclusion 32
• References 33
• 中 文 摘 要 35
• 致 谢 37