Can Twitter be used to Predict Moves in the Stock Market?
A ground breaking study by behavioral scientists has revealed that the mood of the public reflected on social media websites such as Twitter could be used to predict fluctuations in the notoriously unpredictable area of stock market forecasting.
The study suggests that information about public mood garnered from Twitter and other social media sites, using state-of-the-art language analysis tools can help predict spikes and troughs in stock markets.
The scientists involved in the study said that its results, using a complex methodology relying on correlating the fluctuations in the Dow Jones Industrial Average (DJIA) with public mood derived from Twitter, were “strongly indicative of a predictive correlation between measurements of the public mood states from Twitter feeds.”
The paper, spear-headed by Johan Bollen, based at the School of Informatics and Computing at Indiana University, entitled: “Twitter mood predicts the stock market” sates: “One could speculate that the general public is presently as strongly invested in the DJIA as financial experts, and that therefore their mood states will directly affect their investment decisions and thus stock market values.”
The team analysed online Twitter mood data gathered from a three month period that included a number of “socio-cultural events” including the US presidential election and Thanksgiving. This information was analysed by OpinionFinder and Google Profile of Mood States (GPOMS), which looks at six moods: Calm, Alert, Sure, Vital, Kind, and Happy. This “sentiment analysis” was then correlated with daily closing values of the DJIA.
The team found that not only was there a strong correlation particularly between Happy and Calm moods during the US presidential election and Thanksgiving but also that public mood was reflected in DJIA closing values.
For example, the day prior to the election saw a significant drop in Calm levels that pointed to higher levels of anxiety in the public mood. However, on election day, Calm levels increased significantly as did Vital, Happy and Kind scores. Thanksgiving Day, by contrast, pointed to higher Happy levels while the other mood indicators were relatively unaffected.
On these occasions, the researchers found that public mood, particularly levels of Calm were reflected in peaks and troughs in the DJIA.
“The calmness of the public (measured by GPOMS) is thus predictive of the DJIA rather than general levels of positive sentiment,” the study states.
The outcome in the study demonstrates the public mood from Twitter feeds is positively correlated with the moves in the DJIA. Following the release of this paper, the researches have been employed by a hedge fund to build a trading model to test their theory.
The past five years has seen significant improvement in sentiment tracking techniques used to extract information from social media sites like Twitter despite the brevity of the information contained within Tweets. And the team of scientists involved in the study have called for further research into tools to better track public sentiment.
