Sentiment Analysis of Arabic Tweets on the Great March of Return using Machine Learning

Abstract

Social media platforms such as Twitter and Facebook are becoming powerful sources of people’s perception of major events. Most people use social media to express their views on various issues and events and develop their information on a diverse economic, political, technical, social and occurrences related to their life. The overarching aim of this paper is to apply machine learning techniques to extract Arab users’ opinions from 500 Arabic tweets on the Great March of Return rallies in the Gaza strip (Gaza border protests) collected over a two years span from 2018 to 2019. The majority of Sentiment Analysis (SA) studies concentrate on the English language, while other popular languages, such as Arabic, are seldom covered. In addition, on the Internet, publicly accessible Arabic datasets are hardly found. Three Arabic sentiment analysis datasets were used to train and evaluate four machine learning algorithms, namely, Support Vector Machine, Logistic Regression, Decision Tree, and Neural Network. In term of accuracy, logistic regression outperformed the other three algorithms with a percentage of 83%. Application of logistic regression on the sample tweets revealed that 85.8% of the tweets opposed the Great March of Return, whereas 14.2% of the tweets supported it.