关键词: 文本挖掘;支持向量回归;新闻;股市;多元回归

Abstract:

 Text mining and support vector regression techniques were adopted to quantify the impact of financial news on Chinese stock market. Then  the multiple regression of econometrics was used to analyze how the online financial news affected stock market returns systematically. It focused on the impact intensity and time of online news on the stock market, and its impact on listed firms with different scales. Experimental results show that firms with Shenzhen Stock Exchange are more affected by the Internetbased financial news than those of Shanghai Stock Exchange and firms with smaller size tend to have a stronger impact on the movements of Chinese stock market. Our findings include that Internet-based financial news contains hard-to-quantify information of firms, which investors incorporate into stock prices timely.

Key words:  text mining; support vector regression; news; stock market; multiple regression