許永真臺灣大學:資訊工程學研究所黃怡靜Huang, Yi-ChingYi-ChingHuang2010-05-182018-07-052010-05-182018-07-052008U0001-2608200817214600http://ntur.lib.ntu.edu.tw//handle/246246/183598人們使用個人化的使用者描述(personal profile)來呈現自己的興趣與特色,並且取得許多線上的服務。這些使用者描述通常不夠完整,它們只包含了簡單的基本描述,而且只能主觀的呈現使用者自己的想法,無法反映出使用者興趣的動態變化,所以它們無法充分顯現個人的特質。在這篇論文裡,我們提出了附有語意關係的標籤式使用者描述的概念與方法。在概念上,意指我們可以利用所擁有的社群多媒體資料中所附含的標籤,有效的建立符合個人興趣與特質的使用者描述。在方法上,我們使用附有權重的標籤與不同強度的語意關係,來顯現使用者的想法與興趣。常識運算(common sense computing)與共現頻率(co-occurrence)能夠用來計算出不同標籤之間的語意關係。為了突顯出使用者標籤描述的特色與不同面向之間的差異,我們將使用者描述以標籤雲的方式做視覺化的呈現,並且加入了三維度的轉場效果,讓使用者更直覺、更自然的利用這樣的介面去搜尋在社群網路上的資料。People construct personal profiles for self presentation and for obtaining online services. Profiles consisting of simple factual data provide an inadequate description of the individual, as they are often incomplete, mostly subjective and cannot reflect dynamic changes. This thesis explores the idea of ``you-are-what-you-tag'', namely, an individual can be effectively profiled by the tags associated with his/her social media. Specifically, this thesis proposes semantic tag-based profiles, profiles that can be represented as a set of semantically related and weighted tags. The strength of the semantic relationships between these tags are calculated using common sense computing and co-occurrence measurements. Moreover, different views of these profiles are visualized as tag clouds via a 3D switch effect. The proposed approach supports an intuitive and novel interface for people to browse/search through a social web site.Acknowledgments ibstract iiiist of Figures ixist of Tables xihapter 1 Introduction 1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Research Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4 Thesis Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4hapter 2 RelatedWork 5.1 Tagging and Folksonomy . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Design of Tagging Systems . . . . . . . . . . . . . . . . . . . . . . . 6.3 Common Sense Computing . . . . . . . . . . . . . . . . . . . . . . . 8.3.1 Cyc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3.2 WordNet . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3.3 Open Mind Common Sense . . . . . . . . . . . . . . . . . . 10.3.4 ConceptNet . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3.5 Semantic Similarity Analysis . . . . . . . . . . . . . . . . . . 10.4 User Profiling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.5 Tag-based and Social Visualization . . . . . . . . . . . . . . . . . . . 14.5.1 Typical Tag Visualization . . . . . . . . . . . . . . . . . . . . 15.5.2 Tag Orbital . . . . . . . . . . . . . . . . . . . . . . . . . . . 17hapter 3 Tag-based Profile with Semantic Relationship 19.1 Problem Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . 20.2 Proposed Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.2.1 Semantic Tag-based Profile . . . . . . . . . . . . . . . . . . 21.2.2 Tag-based Profile Presentation . . . . . . . . . . . . . . . . . 22hapter 4 Semantic Relationship Analysis 23.1 Three Types of Knowledge . . . . . . . . . . . . . . . . . . . . . . . 24.2 Personal Association: Co-occurrence . . . . . . . . . . . . . . . . . . 25.3 Community Knowledge: Social Wisdom . . . . . . . . . . . . . . . . 27.4 Global Knowledge: Semantic Similarity . . . . . . . . . . . . . . . . 28.4.1 WordNet-based similarity . . . . . . . . . . . . . . . . . . . 28.4.2 ConceptNet-based similarity . . . . . . . . . . . . . . . . . . 29.5 Semantic-based Co-occurrence . . . . . . . . . . . . . . . . . . . . . 32.5.1 Tag Concept Based on Semantic Similarity . . . . . . . . . . 33.5.2 Semantic Co-occurrence Based on Tag Concept . . . . . . . . 34hapter 5 Tag-based Profile Presentation 36.1 Data Characteristic . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.2 Our Idea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.3 Profile Presentation From Three Viewpoints . . . . . . . . . . . . . . 39hapter 6 Experiment and Evaluation 42.1 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.2 User Study Design . . . . . . . . . . . . . . . . . . . . . . . . . . . 44.3 Experiment Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 46hapter 7 Conclusion 49.1 Summary of Contributions . . . . . . . . . . . . . . . . . . . . . . . 50.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51ibliography 52application/pdf3810690 bytesapplication/pdfen-US標註使用者描述社群媒體語意分析語意關聯視覺化呈現tagginguser profilingsocial mediasemantic analysissemantic relationshippresentation賦有語意關聯的視覺化標籤式使用者描述Tag-based Profile Presentation with Semantic Relationshipthesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/183598/1/ntu-97-R95922045-1.pdf