Fuzzy financial time series forecasting models: Problem and Challenge


This talk introduces financial time series forecasting,and discusses fuzzy time series forecasting model and its detailed procedure. Experimental results on our two models (FTSGA and FTSACOAR) are reported. Actual trading data of Taiwan Capitalization Weighted Stock Index is used as benchmark data. Computational results show that our models are efficient, in particular, FTSACOAR outperforms other existing models. Problem and challenge are outlined.


Prof. Defu ZHANG
Professor, Associate Head, Department of Computer Science
Xiamen University, China

Date & Time

26 Sep 2013 (Thursday) 11:00 - 12:00



Organized by

Department of Computer and Information Science


Prof. Defu Zhang received his B.S and M.S degrees from Department of Mathematics of Xiangtan University and his Ph.D. degree from Huazhong University of Science & Technology. He was a senior researcher of Shanghai Jinxin financial engineering academe. Now he works in the department of Computer Science at Xiamen University. He worked as a PostDoc at the Longtop for financial data mining group. He worked as a visiting scholar of Hong Kong City University in summer (08, 09, 11, 12, 13). He worked at University of Wisconsin-Madison as a visiting scholar in 2010. His research interests include all aspects of computational intelligence, data mining. He has published more than 50 papers and developed some packing, timetabling, credit scoring, and stock forecasting software. His ESI(Essential Science Indicators) papers ranks No.1 in his department during past ten years. His research group homepage is His ACM teams obtained 8 silver medals and 3 gold medals since 2005, and entered the final contest in 2007. He is a senior member of China Computer Federation and a member of Computer Theory Professional Committee. He often reviews papers for some international Journals such as INFORMS JOC, EJOR, COR, IEEE TKDE. His projects come from NSFC and companies such as Huawei.