User Profiling for Search Engine Adaptation

PhD Qualifying Examination


Title: "User Profiling for Search Engine Adaptation"

Mr. Wai-Ting Leung


Abstract:

Queries submitted by search engine users tend to be short and ambiguous.
These short queries are unlikely to be able to precisely express what the
users really need. A user profile capturing a user's personal preferences
can help to disambiguate the actual intent of a query. However, users are
usually reluctant to explicitly provide their preferences because of the
extra human effort involved. To alleviate this problem, most recent
research focuses on learning user preferences from user search histories
in an automatic manner. This paper surveys the user profiling methods for
search engine clickthroughs. The methods are classified into three
categories, namely clickthrough-based user profiling, conceptual-based
user profiling and collaborative filtering. Methods in these three
categories are reviewed and compared. Finally, we discuss several possible
research directions for future development of Web search engines.


Date:     		Friday, 2 May 2008

Time:                   2:00p.m.-4:00p.m.

Venue:                  Room 3301A
			lifts 17-18

Committee Members:      Prof. Dik-Lun Lee (Supervisor)
			Dr. Wilfred Ng (Supervisor)
			Dr. James Kwok (Chairperson)
			Dr. Lei Chen
			Prof. Frederick Lochovsky


**** ALL are Welcome ****