Predictive Modeling of Crime Types and Occurrences Using Machine Learning and Deep Learning

This paper review focuses on the potential of applying machine learning algorithms such as K-Nearest Neighbours, Naïve Bayes, Random Forest, and Artificial Neural Network to analyze crime prediction by type and occurrence at specific locations and times and regarding historical data of similar scenarios. From these studies, it is possible to conclude that applying these predictive models can be beneficial when sharing knowledge with law enforcement agencies, particularly regarding optimized resource allocation and strategic security plans. Further research may focus on improving the prediction of new types of threats through stronger ensemble methods, transfer learning between cities, and real-time processing to adapt to changes in crime characteristics. It ultimately helps to reduce the risk to residents and prevent more crime incidents.