New York – Two interesting stories about trading strategies based on sentiment analysis. First, via Bloomberg, a new hedge fund that will use Twitter to predict moves in the stock market:
Derwent Capital Markets, a family- owned hedge fund, will offer investors the chance to use Twitter Inc. posts to gauge the mood of the stockmarket, said co-owner Paul Hawtin.
The Derwent Absolute Return Fund Ltd., set to start trading in February with an initial 25 million pounds ($39 million) under management, will follow posts on the social-networking website. A trading model will highlight when the number of times words on Twitter such as “calm” rise above or below average.
A paper by the University of Manchester and Indiana University published in October said the number of emotional words on Twitter could be used to predict daily moves in the Dow Jones Industrial Average. A change in emotions expressed online would be followed between two and six days later by a move in the index, the researchers said, and this information let them predict its movements with 87.6 percent accuracy.
“Sentiment and mood dramatically change the impact of positive and negative news stories,” said Hawtin in a telephone interview. “If the market’s in a very positive and bullish mood, it can shrug off bad news — bad news comes out and you expect the Dow to fall, and it doesn’t.”
Twitter now has 175 million users and sees 95 million posts per day, according to its website. That has risen from 50 million per day as of February, and researchers are finding new uses for this rapidly growing source of real-time data.
The hedge fund’s strategy seems to be based on research conducted by Johan Bollen and Huina Mao at IU Bloomington and Xiao-Jun Zeng at the University of Manchester. We blogged about this research, which attempts to predict movements in the stock market by tracking public mood as reflected in Twitter posts, using a “Self-Organizing Fuzzy Neural Network” similar to ones already used to successfully forecast electrical load needs. The Bollen/Mao/Zeng paper is available here.
In other news, the New York Times has an article about quantitative sentiment analysis systems which attempt to extract insights into investor sentiment through a semantic analysis of news reports, editorials, company Web sites, blog posts and Twitter messages. The article features the interesting automated sentiment analysis work done by Lexalytics and others, and some of the usual hand-wringing about investment research techniques that seems obligatory in any news article about such things nowadays.