Xiaoou Li

Support vector machine (SVM) is not suitable for classification on large data sets due to its training complexity. Convex hull can simplify SVM training. However, the classification accuracy becomes lower when there are inseparable points. This paper introduces a novel method for SVM classification, called convex-concave hull. After a grid processing, the convex hull is used to find extreme points. Then we detect a concave (non-convex) hull, and use its vertices to train SVM. We finally apply the proposed method to several problems. Experimental results demonstrate that our approach has good classification accuracy while the training is significantly faster than other existing methods.

Biography: Dr. Xiaoou Li obtained B. S degree of applied mathematics in 1991 and PhD degree of Electrical & Electronics Engineering in 1995 from Northeastern University, Shenyang, P. R. China. She had worked in Northeastern University, China for two years after PhD study; then she was a postdoc of National Autonomous University of Mexico (UNAM) from 1998 to 2000. Since April 2000, she has been a professor of Department of Computer Science, The Research and Advanced Studies Centre of the National Polytechnic Institute (CINVESTAV-IPN), Mexico. She was a senior research fellow of School of Electronics, Electrical Engineering & Computer Science, Queen's University Belfast, UK during the school year 2006-2007; and a visiting professor of School of Engineering, University of California Santa Cruz from January to December in 2010. Currently she is a senior member of IEEE (Institute of Electrical and Electronics Engineers), member of AMC (Academia Mexicana de Ciencias, in English: Mexican Academy of Science), and member of SNI (Sistema Nacional de Investigadores, in English: National Researcher System) level 2. Her research interests include knowledge based system, machine learning and data mining applications, social network analysis, Petri nets, non-linear system identification and control, neural networks, system modeling and simulation, active database system, etc.. She has published more than 100 papers in international journals, book chapters and conferences. She has directed 4 doctoral theses and 15 master theses in CINVESTAV. She has successfully finished two CONACYT (NSF in Mexico) projects in the field of Knowledge and Data Engineering and one collaborative project with University of California Riverside.