“…This is because that stock market is noisy, chaotic, nonparametric and non-linear in nature, and many external entities like politics, human psychology/behavior, liquid money and related news influence the direction of the stock market (Abu-Mostafa and Atiya, 1996). Recently, a lot of interesting work has been carrying on in the area of applying machine learning algorithms, including support vector machine (SVM), for analyzing price patterns and predicting stock prices and index changes (Yang et al, 2002;Grosan and Abraham, 2006;Chen et al, 2006;Sapankevych and Sankar, 2009;Kao et al, 2013;Kazem et al, 2013;Zhi-gang et al, 2013). The advantage of SVM is that it is able to reach the global optimum and is resistant to the undertraining or overtraining problems (Yoo et al, 2005;Chen et al, 2006;Sapankevych and Sankar, 2009).…”