This paper proposes a model for the identification of criminal events through the analysis of journalistic news implementing classification mechanism. The classification process is composed of three sub-process: Information Extraction, Classification process and a Selection process of the classes with the best scores obtained after the classification. To obtain the harmonic mean between recall and precision (F-Score) of this classification model, a criminological corpus called CAD was used to simulate different scenarios. CAD is a corpus in Spanish composed of news reporting crimes about homicide, assaults, kidnapping, sexual abuse, and extortion, called High Impact Crimes according to [1].
展开▼