Data Mining for Cyberbullying and Harassment Detection in Arabic Texts

Broadly cyberbullying is viewed as a severe social danger that influences many individuals around the globe, particularly young people and teenagers. The Arabic world has embraced technology continues using it in different ways to communicate inside media platforms. However, text drawbacks for its complexity, challenges, scarcity of resources. This paper investigates several questions related content how protect an from cyberbullying/harassment through information posted on Twitter. To answer this question, we collected Arab corpus covering topics with specific words, which will explain detail. We devised experiments investigated learning approaches. Our results suggest deep models like LSTM achieve better performance compared other traditional classifiers accuracy 72%.