*********************************************************************

                                                              Seminar

Department of Systems Engineering and Engineering Management

                                 The Chinese University of Hong Kong

-------------------------------------------------------------------------------------------------------

Title:   Classification without Providing Negative Examples

Speaker: Dr. Gabriel Pui Cheong Fung

               Department of Systems Engineering and Engineering Management

               The Chinese University of Hong Kong

Date     :   November 22nd, 2006 (Wednesday)

Time    :   4:30p.m. - 5:30p.m.

Venue  :   Room 513

                MMW Engineering Building(Engineering Building Complex Phase 2)

                CUHK

-------------------------------------------------------------------------------------------------------

Abstract:

Traditionally, building a classifier requires two sets of examples:

positive examples and negative examples. Yet, obtaining the negative

examples is sometimes very costly. In this seminar, we will discuss how to

build a text classifier when we are only given some positive examples and

a large set of unlabeled examples. The unlabeled examples are mixed with

both positive examples and negative examples. Since no negative example is

given explicitly, the task of building a reliable text classifier becomes

very challenging. Simply treating all of the unlabeled examples as

negative examples and build a classifier thereafter is undoubtedly a poor

approach to tackle this problem. In this seminar, a labeling heuristics

called PNLH is proposed. It aims at extracting a high quality of positive

examples and negative examples from the unlabeled examples. We will

discuss how PNLH is formulated as well as its impacts on text

classification.

-------------------------------------------------------------------------------------------------------

Biography:

Gabriel Fung received his Ph.D. degree from the Department of Systems

Engineering and Engineering Management at the Chinese University of Hong

Kong in 2006. He is currently an Instructor in the same department from

2005. In 2004, his paper titled "Classifying Text Streams in the Presence

of Concept Drifts" received the best student paper award from the 8th

International Conference of Pacific-Asia Knowledge Discovery and Data

Mining (PAKDD'04). His research interests include data mining, information

retrieval, information systems management and database systems. Gabriel has

published articles in IEEE Transactions on Knowledge and Data Engineering

and IEEE Intelligent Informatics Bulletin, as well as several related

conferences including VLDB, ICDE and ICDM.

***********************  ALL ARE WELCOME  ************************

Enquiries: Peixiang Zhao or Jeffrey Xu Yu,

                 Department of Systems Engineering and Engineering Management

                 CUHK

Website:   http://www.se.cuhk.edu.hk/~seg5810

Email:       seg5810@se.cuhk.edu.hk

********************************************************************