Arabic Cyberbullying Detection: Enhancing Performance by Using Ensemble Machine Learning
Résumé
Cyberbullying is a prevalent threat nearing Arab adolescents and youth around the world. As the internet and smart devices are taking more and more space in the lives of youth, interest in finding solutions for hindering cyberbullying is rising. Interest in Arabic Natural Language Processing techniques is also taking a big part in current research works. A lot of research work nowadays is presenting solutions for the automatic detection of cyberbullying. However, very scarce solutions are prevailing for cyberbullying in Arabic language content. Cyberbullying detection solutions employ Machine Learning and Natural Language Processing techniques. Lately, Ensemble Machine learning techniques had been contemplated as means of enhancing the classification of machine learners. Thus, this paper presents a solution for Arabic Cyberbullying detection. The solution presented hereby uses Ensemble Machine Learning techniques to achieve an enhancement over a previous work presented by the authors in the realm of Arabic Cyberbullying detection. The enhancements are assessed by means of performance measures - that is Precision, Recall and F-Measure.