With the popularity of social networking sites, a lot of information is available on them. Some are authentic while others may or may not be fabricated. It is up to us to figure out the originality of the post. At most we could get a wild guess or with certain details would be able to ascertain the truth behind it. There is a 50% chance of our guess being correct.
Three researchers, Carlos Castillo, Marcelo Mendoza, and Barbara Poblete have developed an algorithm which acts as a Twitter Lie Detector. It can tell about 86% of the time if the information in the tweet is the truth or not. It passes a tweet through a number of tests for the result. If the tweets are long, contains URLs or comes from a well followed source, it may be authentic. If the tweet uses first or third person pronouns, it should not be trusted.
The factors which the programs considered are all those which a human looks for ascertaining the truth. A computer program can go through a pile of tweets in a matter of minutes while humans may take far longer time to evaluate the tweet. The research will be published in the journal Internet Research next month.
