A new stock selection model based on Multi-class Support Vector Machine

Abstract: This paper presents a new stock selection model based on multi-class support vector machine by employing kernel principal component analysis to avoid the risk of speculation and gain the excess return. First, an initial stock pool is choosed according to the industry rotation theory. Then the stock selection index system is constructed based on factor analysis of stock financial indicators and market indicators. Finally, the empirical experiment shows that the proposed stock selection model greatly improves operational efficiency and prediction accuracy for incomplete China's stock market. Keywords: Stock selection model, Kernel Principal Component Analysis , Multi-class Support Vector Machine