Trust-Enhanced Personalized Knowledge Sharing Via Search Inputs, Social Bookmarks and Tags

資訊超載早已是惡名昭彰的問題。尤其在目前以搜尋服務為主力的資訊取得模式下，許多真正被需要的線上資源常常因為受限於個人對於專業領域的掌握程度不足﹝導致想不出合適的搜尋關鍵字﹞而久不見天日。面對這樣的情況，如果系統能夠主動推薦資訊、同時考慮個人化的需求，相信能降低不少前述資訊超載所帶來的衝擊與挑戰。研究提出了運用從社會性標籤(Social Tags)與搜尋關鍵字(Search Keywords)搜集而來的詞彙，為使用者、線上內容/文件以及前述的「詞彙(terms)」(包含社會性標籤或搜尋關鍵字)建構以向量空間模式(VSM)為基礎的識別標誌(Profile)；同時一併保存由搜尋關鍵字與從搜尋結果中去蕪存菁地選為線上書籤(Online Bookmark)的文章，並將這樣的組合視為一種珍貴且值得推薦的資源。而透過使用者與使用者、使用者與文章、使用者詞彙的相關度計算，系統將透過上述的推薦以提供一種個人化知識分享的服務，嘗試更貼近使用者的需求。考慮識別標誌間的「相關度」時，我們除了引用常見的相似度(Similarity)計算，亦延伸了一種對人下標籤的「聯繫人管理(Contact Management)」機制作為具體化使用者間信任之平台以及個人表達需求的管道。對某人下標籤將使該人識別標記中與所被下之標籤相對應的特徵因此被加權，進而影響相關度/相似度的計算使結果更能反應個人的偏好與需求。後，我們亦延伸了常見的「加為好友」社群功能以加權來自於「朋友」的標籤(記錄)，並透過非對襯式的「標籤共存(tag co-occurrence)」分析以提供客製化的標籤關聯以提供更好的個人化服務。Information overload has been notoriously posing tremendous obstacles to more efficient and effective (online) resource utilization. Under the dominance of information pull, in which users have to actively find what they need, services being able to push what we would be interested in or in demand of are often expected.n this paper, we propose an approach, with which social tags and keywords extracted from search inputs are coordinated in terms of VSM to profile users, online contents/documents, and the textual terms themselves. The system stores pairs of users’ search inputs and (Internet) bookmarks selected from the search results, and treats the resulting pairs as valuable resources that are worthy of recommendation. By computing user-user, user-document, and user-tag relatedness values from VSM-based profiles, the system aims to achieve personalized knowledge sharing by recommending the previously-mentioned search input-output pairs that are expected to be related to and catering for users’ needs.n addition to (profile) similarity, we also take interpersonal “trust” into consideration while defining “relateness.” By adopting an innovative contact management mechanism—People Tagging, we allow users to express their preferences for recommendations and their willingness to trust others in specific domains, therefore making recommendations more relevant.astly, based on commonly-seen social network function, for each individual we weight tagging records from people accepted as friends/buddies and provide customized tag relatedness based on an asymmetric tag co-occurrence measure. With these features we expect to achieve higher level of personalization.List of Contents詞 IBSTRACT II要 IIIIST OF CONTENTS 1IST OF TABLES 4 INTRODUCTION 5.1 BACKGROUND—INFORMATION OVERLOAD 5.2 MOTIVATIONS 6.2.1 The Limitations of Contemporary Search Services 7.2.1.1 Search Input Problem 7.2.1.2 Search Output Problem 9.2.2 The Potential of (Social) Tags 10.2.3 The Potential of Search Inputs (From Social Tagging’s Point of View) 11.3 DESIGN PREMISE 12.3.1 Which is the Targeted Information World 13.3.2 What Does the Targeted World Look Like 13.4 APPROACH & CONTRIBUTION 15.4.1 Approach Overview 15.4.2 Proposed Contribution 16.5 THESIS STRUCTURE 17 LITERATURE REVIEW 19.1 FOLKSONOMY 19.1.1 Folksonomy’s Flexibility 19.1.2 Folksonomy’s Richness 21.1.3 Some Examples 23.2 RECOMMENDER SYSTEM (RS) 25.2.1 Content-Based Filtering (CBF) 26.2.1.1 Pros and Cons 27.2.2 Collaboratibe Filtering 28.2.2.1 User-based CF 28.2.2.2 Item-based CF 29.2.3 Summary of RS Algorithms 31.2.4 Tag Recommendation 32.2.5 Search Input Recommendation 33.3 TRUST 35.3.1 About Trust 35.3.1.1 Scopes of Trust 36.3.1.2 Trust V.S. Reputation 36.3.1.3 Trust’s Characteristics 37.3.2 Trust-enhanced Recommender System 38.3.2.1 Moleskiing [53,54] 38.3.2.2 FilmTrust [55] 39.4 SUPPLEMENTARY CONCEPTS 40.4.1 Concept Space 40.4.2 Concepts of Vector Profiles & Similarity Computation 41.4.2.1 For Tags 41.4.2.2 For documents and people 42.4.3 Profile Construction Methods 43.4.3.1 Naïve Approach 43.4.3.2 Co-occurrence Approach 45.4.3.3 Adaptive Approach 46.4.4 People Tagging 46 SYSTEM DESIGN 48.1 DESIGN GOAL 48.2 SYSTEM CONCEPT 48.2.1 Proceeding Activities—Data Collecting 49.2.2 (User-based) Collaborative Filtering 50.2.3 Content-based Filtering 51.2.4 Trust in Action 52.2.4.1 People Tagging 52.2.4.2 BuddyList 54.2.5 Extracting Terms’ (Inter-)relatedness 55.2.5.1 Asymmetric Tag Co-occurrence Measure 55.2.5.2 (Term/Domain) Profile Weighting 56.3 SYSTEM ARCHITECTURE 57.3.1 Use Case Diagram 58.3.2 Sequence Diagrams 59.3.2.1 Query Recommendation 59.3.2.2 People Tagging 61.3.2.2.1 People Tag Collecting 61.3.2.2.2 People Tag Utilizing (6~12) 62.3.2.3 BuddyList 64 SYSTEM IMPLEMENTATION 66.1 DEVELOPMENT ENVIRONMENT 66.2 SYSTEM FEATURES 66.2.1 Keyword Extraction 66.2.2 Typical Social-Bookmarking-Website-Like Functions 69.2.3 User Profile 70.2.4 People Tagging and Search Input Recommendation 71.2.4.1 Implicit Self-Tagging 73.2.4.2 Query-bookmark Pair’s Forming Scenarios 74.2.5 Updating BuddyList 76.2.6 Summary of Inter-PeopleTagging 77.2.7 Contextual Information 77.2.8 Expert Generation 78.3 EMPIRICAL EVALUATION OF OUR PERFORMANCE 79.4 DATABASE SCHEMA 82 CONCLUSION AND DISCUSSIONS 85.1 SUMMARY 85.2 CONTRIBUTIONS 86.2.1 From Functional View 86.2.2 Compare and Contrast 88.3 LIMITIONS 91.4 FUTURE WORKS 92 BIBLIOGRAPHY 9