Proposed Deep Learning Model for Rumour Detection in Facebook Posts in Arabic Language

Abstract

The emergence of Social Network Sites (SNSs) has led the individuals in general to find easy way for rapid communication with each other at any time and place. Information that spread out through (SNSs) can include a lot of unreal allegations, in which rumours and fake news on some specific manner can proliferate readily causing to a vast amount of problems. This paper addressed in order to detect the rumour posts in Facebook social site in Arabic language. The proposed work mainly relies on using sentiment analysis in order to prepare the data for extracting useful features. The Deep Convolutional Neural Network (CNN) classification model is proposed and adopted in order to perform classification operation for the extracted features from the Arabic posts. The experiential results were magnificent in which all of accuracy, precision, f-measure, and recall were equal to100%.