« Seminar - Design Automation of Configurable Processors | Main | Information Security Summit 2006 Hong Kong »
November 01, 2006
Seminar - Feature Selection for Pairwise Scoring Kernels with Applications to Protein Subcellular Localization
Speaker:
Prof. S.Y. Kung, PhD, FIEEE
Professor, Department of Electrical Engineering
Princeton University, U.S.A.
Date:
2 Nov. 2006 (Thur.)
Time:
5:00 p.m. - 6:00 p.m.
Venue:
CYC603, 6/F, Chow Yei Ching Building,
Dept. of Electrical and Electronic Engineering,
The University of Hong Kong
Abstract:
Pairwise scoring kernels have been used extensively in biological sequence classification because of their effectiveness in converting variable-length sequences into fixed-length vectors. However, the pairwise approach can result in feature vectors with dimension equal to the training set size, causing the curse of dimensionality. This difficulty calls for feature selection methods that can weed out irrelevant features to reduce training and recognition time. To this end, we propose to use the full-feature column vectors of a pairwise scoring matrix to train an SVM and select the feature dimensions (rows) based on its support vectors. The idea is based on the notion that support vectors are important for classification and pairwise scoring matrices are symmetric. As a result, the transpose of support vectors can be considered as important features for classification. We refer to this approach as vector-index-adaptive SVM (VIA-SVM) and compare its performance with other feature selection schemes---including SVM-RFE, R-SVM, and a Fisher-based method---in protein subcellular localization. It was found that VIA-SVM is insensitive to the penalty factor in SVM training and can avoid the need to set a cutoff point for stopping the feature selection process.
About the Speaker:
Sun-Yuan Kung was born in Taiwan on January 2, 1950. He received the B.S. in Electrical Engineering from the National Taiwan University in 1971; M.S. in Electrical Engineering from the University of Rochester in 1974; and Ph.D. in Electrical Engineering from Stanford University in 1977. From 1977 to 1987, he was on the faculty of Electrical Engineering-Systems at the University of Southern California. In 1984, he was a Visiting Professor at Stanford University and later in the same year, a visiting professor at the Delft University of Technology. Since September 1987, he has been a Professor in the Department of Electrical Engineering, Princeton University. He currently serves on the IEEE Technical Committees on VLSI Signal Processing and Neural Networks and an Editor-in-Chief of Journal of VLSI Signal Processing.
Enquiries:
Please contact Dr. C.Q. Chang, Department of Electrical and Electronic Engineering, The University of Hong Kong (Email: cqchang@eee.hku.hk)
Posted by ymlam at November 1, 2006 11:45 AM