Social Unrest Prediction Through Sentiment Analysis on Twitter Using Support Vector Machine: Experimental Study on Nigeria’s #EndSARS

Abstract

Abstract Social unrest is a powerful mode of expression and organized form of behavior involving civil disorders and acts of mass civil disobedience, among other behaviors. Nowadays, signs of most social unrest start from the social media websites, such as Twitter, Facebook, etc. In recent times, Nigeria has faced different forms of social unrest, including the popular #EndSARS, which began on Twitter with a demand that government disband the Special Anti-Robbery Squad (SARS), a unit under the Nigerian Police Force for alleged brutality. Mining public opinions such as this on social media can assist the government and other concerned organizations by serving as an early warning system. In this work, we collected user tweets with #EndSARS from Twitter and pre-processed and annotated them into positive and negative classes. A support vector classifier was then used for classifying the sentiment expressed in them. Experimental results show 90% accuracy, 94% precision, 85% recall, and 89% F 1 score on the test set. The codes and dataset are publicly available for research use. https://github.com/Temidayomichael/Social-unrest-prediction .