Scientists Develop Algorithms to Distinguish Truthful Tweets From Lies

December 17, 2012 - 12:15 PM

A team of scientists have developed algorithms to sort out the truthful Tweets from the lies, according to a report that will be published next month in Internet Research.

The algorithm uses a series of tests to predict the truthfulness of information posted on Twitter. Some of the input that is examined includes determining if messages come from a well-known source, are longer in length, or contain URLs (links).

The test also studies the language that is used: Question marks, exclamation marks and first or third-person pronouns are all signs that a tweet isn't truthful.

The algorithm can analyze "mountains" of tweets in minutes and is said to have an 86 percent accuracy rate in determining a Tweet's truthfulness.

Twitter is a social networking and microblogging site that was founded in 2006.  Since it's inception it has added over 500 million users and generates more than 340 million Tweets per day.

Determining the veracity of social media content has become a daunting challenge as more people have started looking to social media as a source of news and information.

See more "Right Views, Right Now."