Machine Learning Classification Algorithms for Rumours Detection in Facebook Arabic Posts

Abstract

In the last few years, the online social network sites have enabled the people around the world to access and share information at tremendous speed from anywhere which makes a lot of them adopted it as the main source of information. The information is circulated in the public domain without adequate awareness and confirmation to support its legitimacy making rumours and fake news increased rapidly. In this work the rumours are detected in the Facebook social site posts that are written in Arabic language. Basically, the post will be classified into real or rumour using machine learning classification algorithms after performing pre-processing, and feature extraction operations. The obtained results acquired from the utilized five machine learning classification algorithms were uneven within high and low accuracy results.