Abstract—Of late there has been a significant amount of work on using sources of text data from the Web (such as Twitter or Google Trends) to predict financial and economic variables of interest. Much of this work has relied on some form or other of superficial sentiment analysis to represent the text. In this work we present a novel approach to predicting economic variables using sentiment composition over text streams of Web data. We treat each text stream as a separate sentiment source with its own predictive distribution. We then use a Bayesian classifier combination model to combine the separate predictions into a single optimal prediction for the Nonfarm Payroll index, a primary economic indicator. Our results show that we can achieve high predictive accuracy using sentiment over big text streams.

Index Terms—Economic prediction, text sentiment, big data streams, Bayesian classifier combination.

Abby Levenberg is with the Oxford-Man Institute of Quantitative Finance and Dept. of Computer Science, University of Oxford, UK (e-mail: abby.levenberg@oxford-man.ox.ac.uk).
Stephen Pulman is with the Dept. of Computer Science, University of Oxford, UK (e-mail: stephen.pulman@cs.ox.ac.uk).
Karo Moilanen is with the TheySay Analytics Ltd.
Edwin Simpson is with the Dept. of Engineering Science, University of Oxford.
Stephen Roberts is with the Oxford Man Institute of Quantitative Finance and Dept. of Engineering Science, University of Oxford.

Cite:Abby Levenberg, Stephen Pulman, Karo Moilanen, Edwin Simpson, and Stephen Roberts, "Predicting Economic Indicators from Web Text Using Sentiment Composition," International Journal of Computer and Communication Engineering vol. 3, no. 2, pp. 109-115, 2014.