Epistemology-based social search for exploratory information seeking

• Main Author: Mao, Yuqing
• Other Authors: Shen Haifeng
• Format: Theses and Dissertations
• Language: English
• Published: 2012
• Subjects: DRNTU::Engineering::Computer science and engineering::Information systems::Information storage and retrieval
• Online Access: https://hdl.handle.net/10356/50765
• Institution: Nanyang Technological University

As the Internet and communication technologies become ever more pervasive, we see an astounding number of new information seeking behaviors. People often start with some vague information need and iteratively seek and select bits of information that cause the data needs and behavior to evolve over time. This phenomenon is referred to as exploratory information seeking, which is open-ended, persistent, and multifaceted. In general, formulating proper keywords and evaluating search results are common difficulties in exploratory information seeking. One possible solution is to utilize social cues provided by a large number of users. The main objective of this thesis is to improve exploratory information seeking by social search, i.e., utilizing the wisdom of crowds. The main contribution of this thesis research is an epistemology-based social search framework, where search epistemologies – aggregated and well-structured information packages derived from successful search processes contributed by numerous searchers – are effectively shared, reused, and refined by others with same or similar search interests. This framework contains several distinctive components that are well-organized and work in cooperation with each other. The first component is the epistemology generation component, where the search epistemologies can be automatically derived from successful search processes of massive users doing exploratory information seeking. The generative process of an epistemology is modeled using a probabilistic topic model with social tags. An approach for query reformulation and results ranking is proposed based on this model. The second component is the epistemology search component, where the search epistemologies can be retrieved for users doing exploratory information seeking with diverse requirements. A social-interest-based query suggestion and results diversification approach is proposed to support exploratory information seeking, while the social interest is discovered by employing the kernel principal component analysis on the related queries and results. The third component is the epistemology editing and refining component, where the search epistemologies can be collaboratively edited in a consumer-led interactive search process. An information provision on demand approach is proposed to help information consumers acquire non-existent information, while invited information providers from relevant social networks jointly refine the epistemologies to meet the consumer’s needs. The fourth component is the epistemology services component, which makes the social epistemology-based search systems viable, reliable, and sustainable. A trust model for trust inference through credit flow in the epistemology-based social search network is proposed, which allows personalized measures derived from trust propagation to be naturally established on the objective ground. A non-monetary incentive mechanism for encouraging users to contribute their epistemologies is proposed, where a knowledge bartering process can be automated through the online silk road that maximizes the social welfare within a social search community. Furthermore, experimental results based on the prototype system Baijia demonstrate these methods can enhance the epistemology-based social search and improve exploratory information seeking with better performance.