Persian Rumor Detection on Twitter

Nowadays, one of the common ways of news broadcasting is sharing news through online social media. Some users consider these social media as a platform for news broadcasting. Every day numerous news transfer among users. However, sometimes rumors get around between users, such that they may make some mistakes about what are exactly happened. If rumors has been recognized at the right time, their negative effects can be bounded. In order to differentiate between rumors and non-rumors tweets, various well-known machine learning methods are applied on KNTUPT dataset which is collected all persian tweets from November 24th, 2017 to December 8th, 2017. The results indicate that the Random forest and meta.RandomSubSpace show their superiority than other methods.