Data-based automatic Covid-19 rumors detection in social networks

Abstract

Social media is one of the largest sources of propagating information, however, it is also a home ground for rumors and misinformation. The recent extraordinary event in 2019, the COVID-19 global pandemic, has spurred a web of misinformation due to its sudden rise and global widespread. False rumors can be very dangerous, therefore, there is a need to tackle the problem of detecting and mitigating false rumors. In this paper, we propose a framework to automatically detect rumor on the individual and network level. We analyzed a large dataset to evaluate different machine learning models. We discovered how all our methods used contributed positively to the precision score but at the expense of higher runtime. The results contributed greatly to the classification of individual tweets as the dataset for the classification task was updated continuously, thereby increasing the number of training examples hourly.