Statistical Relational Classification of Networked Data

PhD Qualifying Examination


Title: "Statistical Relational Classification of Networked Data"

Mr. Wujun Li


Abstract:

To simplify the learning models, traditional machine learning
methods assume that instances are independent and identically
distributed (i.i.d). However, most real-world data are
relational in the sense that different instances are related
to each other. Networked data are a special type of
relational data in which the instances are interconnected, such as
web pages. In networked data, the attributes of connected (linked)
instances are often correlated and the class label of one instance
may have an influence on the class label of a linked instance.
Hence, naively applying traditional learning methods to networked
data may lead to misleading conclusion about the data. Because
networked data widely exist in a lot of application areas, such as
web mining, social network analysis, bioinformatics, and marketing
and so on, recently many researchers have started to propose novel
methods, called statistical relational learning(SRL) methods,
to model the networked data. With the focus on the
classification methods which try to classify the instances in
the networks, we review this class of methods, called
statistical relational classification (SRC) methods, for
networked data in this article. Furthermore, some possible research
directions are also pointed out based on a comprehensive analysis of
the existing SRC methods.


Date:     		Monday, 21 January 2008

Time:                   3:00p.m.-5:00p.m.

Venue:                  Room 3304
			lifts 17-18

Committee Members:      Dr. Dit-Yan Yeung (Supervisor)
			Prof. Qiang Yang (Chairperson)
			Dr. Brian Mak
			Dr. Weichuan Yu (ECE)


**** ALL are Welcome ****