Terrorist Act Prediction Based on Machine Learning: Case Study of Tunisia

Over the last years, considerable literature has grown up around the theme of the fight against terrorism. In fact, even during the COVID-19 pandemic, terrorist acts have been carried out in several countries causing several deaths and injuries. We propose in this work to predict the main attributes related to an act of terrorism, such as the place, the suspect terrorist group, the attack type, and the victim. We study particularly the case of Tunisia by exploiting the two well known databases on international terrorism: GTD and ACLED. Four machine learning algorithms are proposed: SVM, decision tree, random forest, and KNN.