Researchers Successfully Predict Stock Market… By Analyzing 10 Million Tweets

In an incredible study at UI Bloomington's School of Informatics and Computing, researchers found that they were able to predict the movement of the stock market based on the average 'mood' of people using 10 million tweets.

In an incredible study at UI Bloomington’s School of Informatics and Computing, researchers found that they were able to predict the movement of the stock market based on the average ‘mood’ of people using 10 million tweets.

The study used two mood-tracking tools to analyze text in various Twitter feeds and determine whether people, in general were feeling calm, alert, sure, vital, kind or happy. They then measured this every day for 10 months during 2008 and found that 90% of the time, there was a correlation between people’s moods.

Specifically, the researchers found that by using a neural network prediction algorithm, and including the public mood data, they were able to increase the accuracy of their predictions of upwards or downwards movement of the Dow Jones Industrial Average each day.

The researchers were using this Tweet study as a first attempt to draw correlations between market values and public ‘mood’ data, and the success of the test means they will be creating more detailed research in the future. We’ll keep you posted.

Recommended articles