PhD Thesis Proposal Defence


"Web Query Classification"

by

Mr. Dou Shen


Abstract:

Web query classification (QC) aims to classify Web users' queries, which
are often short and ambiguous, into a set of target categories. QC has
many applications including improving page ranking inWeb search, targeted
advertisement in response to search queries, and personalization. However,
the sparsity and ambiguity of Web queries renders QC fundamentally
diRerent from traditional document classification. In this proposal, we
present two solutions which can solve the QC problem eRectively and a
third potential solution. In the first solution, we enrich both Web
queries and target categories through Web search engines and existing
taxonomies, based on which we develop two kinds of classifiers as well as
an ensemble classifier. The second solution improves thefirst one by
introducing a bridging classifier, which takes some existing taxonomies to
bridge Web queries and target categories. The third solution, which is
applicable to solve the general data sparsity problem in text mining,
calculates term relationships from existing text corpus. After that, we
augment Naive Bayes classifiers with the discovered term relationships.
Preliminary results over an open benchmark dataset from KDDCUP 2005
validate the eRectiveness of the first two solutions. In the future, we
hope to further study the first two solutions and test the third solution
empirically.


Date:     		Monday, 12 February 2007

Time:                   10:00a.m.-12:00noon

Venue:                  Room 3501
			lifts 25-26

Committee Members:      Dr. Qiang Yang (Supervisor)
                        Dr. Brian Mak (Chairperson)
			Dr. Lei Chen
			Dr. James Kwok


**** ALL are Welcome ****