With the widespread of modern technologies and social media networks, a new form of bullying occurringanytime and anywhere has emerged. This new phenomenon, known as cyberaggression or cyberbullying, refersto aggressive and intentional acts aiming at repeatedly causing harm to other person involving rude, insulting,offensive, teasing or demoralising comments through online social media. As these aggressions represent a threat-ening experience to Internet users, especially kids and teens who are still shaping their identities, social relationsand well-being, it is crucial to understand how cyberbullying occurs to prevent it from escalating. Considering the massive information on the Web, the developing of intelligent techniques for automatically detecting harm-ful content is gaining importance, allowing the monitoring of large-scale social media and the early detection ofunwanted and aggressive situations. Even though several approaches have been developed over the last few yearsbased both on traditional and deep learning techniques, several concerns arise over the duplication of research andthe difficulty of comparing results. Moreover, there is no agreement regarding neither which type of technique isbetter suited for the task, nor the type of features in which learning should be based. The goal of this work is toshed some light on the effects of learning paradigms and feature engineering approaches for detecting aggressionsin social media texts. In this context, this work provides an evaluation of diverse traditional and deep learningtechniques based on diverse sets of features, across multiple social media sites.