Predicting Sector Index Movement with Microblogging Public Mood Time Series on Social Issues

Predicting Sector Index Movement with Microblogging Public Mood Time Series on Social Issues Yujie Lu1 Jinlong Guo2 Sakamoto Kotaro1 Shibuki Hideyuki1 Tatsunori Mori1 1Graduate School of Environment and Information Science, Yokohama National University, Yokohama, 2408501, JAPAN 2Graduate School of Library and Information Science, University of Illinois at Urban-Champaign, Champaign, IL 61820, USA {luyujie,sakamoto,shib,mori}@forest.eis.ynu.ac.jp, jguo24@illinois.edu Abstract This paper develops a technique that unfolds public mood on social issues from real-time so- cial media for sector index prediction. We first propose a low-dimensional support vector ma- chine (SVM) classifier using surrounding infor- mation for twitter sentiment classification. Then, we generate public mood time series by aggre- gating message-level weighted daily mood (WDM) based on the sentiment classification re- sults. Lastly, we evaluate our method against the real stock index in two kinds of time periods (fluctuating and monotonous) separately using static cross-correlation coefficient (CCF) and dynamic vector auto-regression (VAR). The ex- periments on “food safety” issue show that the proposed WDM method outperforms the word- level baseline method in predicting stock move- ment, especially during fluctuating period. 1 Introduction Social media websites, such as Twitter and Face- book, have generated a great amount of public opin- ions on a variety of issues, especially hot events and emergencies. As a result, user-generated content has become a significant resource for exploring useful knowledge. In the use of public mood entailed in the real-time message streaming, researchers have pro- posed a wide range of applications, for example, election prediction (Andranik et al., 2009), anti-ter- rorism assistance (Cheong and Lee, 2009) and con- sumer confidence poll (O’Connor et al., 2010). In this paper, we use it for stock prediction. 1 According to 2015 first quarter financial results released by Weibo Corp. Sina Weibo is a Twitter-like microblogging service in China. Launched in 2009, it now has near 200 million monthly active users1, which makes it the most dominant social networking service in China. Users discuss all kinds of social topics and express their opinions on the platform. As an example, food safety issue has become a prominent social problem and caused much concern in recent years in China. Figure 1 is an example tweet talking about food safety from Sina Weibo, in which the author ex- presses his dissatisfaction to the situation of food safety in China. Note that besides the text part, there is auxiliary information around the text (called sur- rounding information in this paper). Previous work shows that indicators from real-time media can conceivably be used to predict changes for many economic indexes (Bollen et al. 2011), and behavioral finance theory suggests that public mood can drive stock market (Nofsinger, 2005).Hence, we construct public mood time series by analyzing mil- lions of tweets in a time span to predict stock move- ment in the corresponding period. Our main contributions are summarized as follows:  We investigate how microblogging public mood on certain social issues relates to the stock move- ment of the relevant sector. In this study, we con- duct an experiment on …

Content truncated to 500 words