Tracking Multiple Social Media for Stock Market Event Prediction
Fang Jin, Wei Wang, Prithwish Chakraborty, Nathan Self, Feng Chen, Naren Ramakrishnan
The problem of modeling the continuously changing trends in finance markets and generating real-time, meaningful predictions about significant changes in those markets has drawn considerable interest from economists and data scientists alike. In addition to traditional market indicators, growth of varied social media has enabled economists to leverage micro- and real-time indicators about factors possibly influencing the market, such as public emotion, anticipations and behaviors. We propose several specific market related features that can be mined from varied sources such as news, Google search volumes and Twitter. We further investigate the correlation between these features and financial market fluctuations. In this paper, we present a Delta Naive Bayes (DNB) approach to generate prediction about financial markets. We present a detailed prospective analysis of prediction accuracy generated from multiple, combined sources with those generated from a single source. We find that multi-source predictions consistently outperform single-source predictions, even though with some limitations.
Fang Jin, Wei Wang, Prithwish Chakraborty, Nathan Self, Feng Chen, Naren Ramakrishnan: Tracking Multiple Social Media for Stock Market Event Prediction. ICDM 2017: 16-30
Professor of Computer Science
- Date of publication:
- July 1, 2017