Rumor Verification
The Korea Advanced Institute of Science and Technology (KAIST) announced on January 9 that a research team led by professor Cha Mee-young at KAIST successfully developed a technology that can determine whether or not rumors on social networking services are true, with 90 percent accuracy. The study was conducted in partnership with Seoul National University and Microsoft Research Asia.
The team analyzed characteristics of more than 100 rumors circulating on Twitter between 2006 and 2009 about politics, IT, health, celebrities, and other fields. The results showed that the method was especially effective in differentiating defamatory rumors that target certain persons and institutions from true stories.
For example, presidential candidate Barack Obama was rumored to be Muslim, anti-Christian, and to have obtained US citizenship illegally. However, those rumors turned out to be false. In addition, the research team found malicious rumors and false information about actress Nicole Kidman, which said that she underwent a sex change operation or she is a bisexual.
Professor Cha said, “This research is done through the convergence of statistics and mathematics models with sociological and psychological theories. After analyzing data, our team was able to grasp the attributes of rumors that tend to receive more attention.” She added, “It is difficult to identify the authenticity of rumors at the very beginning. But the identification is made possible after time passes, or when those rumors are widespread, based on a sufficient amount of data.”
The research findings were presented at the IEEE International Conference on Data Mining in December 2013.