Prediction Web Application Based on a Machine Learning Model to Reduce Robberies and Thefts Rate in Los Olivos, San Martín de Porres and Comas

Sanchez M.; Estefano L.; Padilla Z.; Antonio C.; Castañeda P.S. · International Conference on Web Information Systems and Technologies, WEBIST - Proceedings (2024) · DOI: 10.5220/0012906800003825

Resumen

Robberies and thefts in the districts of Los Olivos, San Martin de Porres and Comas in Lima, Peru are a constant problem. The scarce police presence on the streets makes these areas ripe for crime. This project proposes analyze crime rates across the public authorities to take measures that might reduce the crime rate with the development of a Machine Learning model, through the use of Random Forest (RF) and a dataset with information from districts in similar situations to those raised in the project. The proposed solution includes a web application interface for data input and analysis, that will be used by municipal entities and everyone. Performance metrics such as Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE) were included, with results showing MAEs of 29.194, 45.219, and 75.572 and RMSEs of 39.651, 58.199, and 93.110 from other districts with the same condition. The study concludes with a refinement of machine learning methodologies for crime prediction and emphasizes the potential for citizen engagement in crime prevention. Copyright

Instituciones: Universidad Peruana de Ciencias Aplicadas