A criminal macrocause classification model: An enhancement for violent crime analysis considering an unbalanced dataset

This study introduces a novel model designed to classify macrocauses of violent crimes. The model's practical application is demonstrated through its integration into the framework of the Natal Smart City Initiative in Brazil. Utilizing the Design Science methodology, the study details the model's development, its subsequent implementation through a machine learning pipeline, and its assessment employing four prominent classification techniques: Decision Trees, Logistic Regression, Random Forest, and XGBoost. XGBoost performed exceptionally well, achieving an average accuracy of 0.961791, an F1-Score of 0.961410, and an AUC of ROC curve of 0.994732. Accurate classification of criminal macrocauses is crucial for developing effective public safety policies. The proposed model can provide public safety institutions and criminal analysts with a valuable tool for better understanding aspects related to violent crime analysis in their cities. This can streamline the analysis and management process and provide more accurate information for decision-making. This study also has important implications for the emerging field of smart cities. By providing a tool to assist in decision-making and planning public safety strategies, this work contributes to the development of innovative, data-based, and theory-based solutions to address urban challenges.