id	author	title	date	pages	extension	mime	words	sentence	flesch	summary	cache	txt
ajfa-5485	Liu, Xuan; Pan, Liu	Prediction of the Moving Direction of Google Inc. Stock Price Using Support Vector Classification and Regression	2014	14	.pdf	application/pdf	4054	176	56	b 2.2 Kernel and optimization of kernel parameters Epsilon-SVR in LIBSVM package, which is developed by Chang et al (2011) and is currently one of the most widely used SVM, was employed for stock price prediction in this study. However, this does not mean that moving trends >15 trading days will always be good for stock price prediction, as we also found that the trends within 45 trading days had the similar prediction performance as that within 15 trading days.	cache/ajfa-5485.pdf	txt/ajfa-5485.txt