Predicting the threat of death in stalking cases through Artificial Neural Network

Abstract

Stalking is a complex phenomenon that needs to be explained through several frameworks of research. During the years, scientists from the psychological, criminological, psychological, and legal fields made important steps to understand the nature of Stalking to find how to stop this type of behavior. The new concept that this research introduces is the possibility to predict if there will be the threat of death against a victim by a stalker through Artificial Intelligence and, more specifically, using an Artificial Neural Network (ANN). The thesis analyses variables impact on the learning process and the accuracy of this new system about predicting the threat of death. Publicly available and collected into the database Violence and Threats of Violence Against Women and Men in the United States built by Tjaden and Thoennes, the secondary data have been analyzed through the ANN. The result showed that variables impact fully support the theories, and the prediction of the threat of death against the victim is possible in this simulated environment with real data.