Forget about insider trading. Use Twitter.
German economists who used machine intelligence to analyze tweets about stocks say the advice shared among Twitter members can accurately predicts what the market will do a day ahead of time.
The key is to pick the right group to listen to among all those tossing out comments about stocks via Twitter, says a research headed by Timm Sprenger, an economist at the Technical University of Munich who is founder of tweettrader.net, a website that analyzes and aggregates relevant tweets.
SOCIAL NETWORK DANGER: Attackers favor social-networking sites, shortened URLs
Their contention is that the stock market efficiently aggregates information and opinions in stocks that are reflected in stock prices and trading volumes. They say proper analysis of tweets about stocks can do the same but a day ahead of the market. They found their results so compelling that they've created the website.
Financial professionals could use this type of analysis to figure out what tweeted information reflects actual investor behavior, the researchers say.
Using Web crawlers plugged into Twitter's API to find mention of S&P 100 stocks, the team used open source Weka software to automatically read and analyze the tweets they gathered, classifying them as buy, hold or sell. It turned out about half were hold, about 35% were buy and about 15% were sell.
They took into account how many times comments were retweeted, how many times they were mentioned by other Twitter members and how many followers individual tweeters had -- all as a way to give weight to individual tweets that are in line with their influence in the Twitter community.
Overall, the most respected tweets are those that get the most mention or are retweeted the most, the researchers say. "Users who provide above average investment advice are retweeted (i.e., quoted) more often, have more followers and are thus given a greater share of voice in microblogging forums," they say in their research paper on the subject.
They looked at 249,533 tweets about stocks listed on the S&P 100 and found a strong correlation between trading volume for a particular stock and the number of tweets that mention it.
The researchers found that tweeters offered up opinions contrary to what the market is currently doing and that the number of tweets about a particular stock is a good prediction of how many shares of it will be traded the next day.
The team looked at financial tweets between Jan. 1 and June 30, 2010.
Read more about lans and routers in Network World's LANs & Routers section.