Title :
Pattern Recognition Based on Support Vector Machine: Computerizing Expertise for Predicting the Trend of Stock Market
Author_Institution :
Manage. Sch., Northwestern Polytech. Univ., Xi´´an, China
fDate :
March 31 2009-April 2 2009
Abstract :
This paper presents a method for forecasting the moving direction of Shanghai Stock Composite Index (CCI) through constructing the feature pattern vectors containing the characters of the market structure according to the profitunity approach and adopting SVM to perform pattern recognition. First, the market trend forecasting is considered as a pattern recognition problem. Second, the pattern vectors are formed with the help of the investing expertise, the profitunity approach which devises a set of rules of predicting market moving trend by observing some variables (designed from the market data) and their combinations. Then, the support vector machine is employed to perform the pattern recognition, mapping the pattern vectors into class space of trend moving up and down. Finally, a group of simulations are given, and the results show the good performance that the correct rate of forecasting reached about 70%.
Keywords :
financial data processing; stock markets; support vector machines; forecasting; market trend; pattern recognition; stock market; support vector machine; Computer science; Economic forecasting; Engineering management; Pattern recognition; Predictive models; Risk management; Statistical learning; Stock markets; Support vector machine classification; Support vector machines; Pattern Recognition; Predicting; Support Vector Machine; Trend of Stock Market;
Conference_Titel :
Computer Science and Information Engineering, 2009 WRI World Congress on
Conference_Location :
Los Angeles, CA
Print_ISBN :
978-0-7695-3507-4
DOI :
10.1109/CSIE.2009.798
