Abstract

Effective policing strategies for mitigating and preventing crimes are currently a significant challenge for law enforcement in the Philippines. While police patrolling helps reduce crime risks through visibility, limitations on resources and technology in police departments heavily hinder the efficiency of traditional crime management. To address operational challenges in resource allocation for crime prevention and response, this study developed an integrated system, “PulisAI”, utilizing machine learning for multi-level crime alarm (low, medium, high) prediction for eight high-priority focus crimes. A local crime dataset, alongside publicly available demographic and geographic information, was put into a comprehensive data preprocessing pipeline. Crime incidents were aggregated into unique spatio-temporal blocks, followed by feature engineering to derive the three-tier alarm level target variable and historical crime metrics. A comparative analysis of four machine learning models—Multinomial Logistic Regression, Random Forest, Support Vector Classifier, and XGBoost Classifier—was conducted under imbalanced data and SMOTE-balanced training conditions. The results showed that the XGBoost Classifier, trained on the original imbalanced dataset, is the superior model, achieving an accuracy of 92.69%, a macro-averaged F1-Score of 87.00%, and an Unweighted Average Recall of 82.00%. The selected model was integrated into the PulisAI web application. PulisAI served as a data-driven tool for decision-making for law enforcement in patrol planning, resources allocation, and crime pattern analysis.

References

• Abubakar, H., Boukari, S., Gital, A. Y., & Zambuk, F. U. (2024). A hyper-parameter tuned Random Forest algorithm-based on Artificial Bee Colony for improving accuracy, precision and interpretability of crime prediction. Dutse Journal of Pure and Applied Sciences, 10(4a), Article 4a. https://doi.org/10.4314/dujopas.v10i4a.34
• Annang, A., Abrokwah, R. K., & Asante, E. O. (2023). Challenges of Effective Policing in Ghana: The Central Regional Command in Perspective. Open Access Library Journal, 10(8), 1–18. https://doi.org/10.4236/oalib.1110506
• Asor, J. R. (2020). Implementation of Predictive Crime Analytics in Municipal Crime Management System in Calauan, Laguna, Philippines. International Journal of Advanced Trends in Computer Science and Engineering, 9(1.3), 150–157. https://doi.org/10.30534/ijatcse/2020/2291.32020
• Ayeng, R. (2024, September 15). Manila tops cities with highest crime rate in SE Asia-Global poll. Daily Tribune. https://tribune.net.ph/
• Balahadia, F. F., Asor, J. R., Catedrilla, G. M. B., Villarica, M., & Cabiente, J. M. (2020). Intelligent Investigation on Crime Incident Reports in the Province of Laguna through Predictive Model Development. International Journal of Advanced Trends in Computer Science and Engineering, 9(1.3), 139–144. https://doi.org/10.30534/ijatcse/2020/2091.32020
• Butt, U. M., Letchmunan, S., Hassan, F. H., Ali, M., Baqir, A., & Sherazi, H. H. R. (2020). Spatio-Temporal Crime HotSpot Detection and Prediction: A Systematic Literature Review. IEEE Access, 8, 166553–166574. https://doi.org/10.1109/ACCESS.2020.3022808
• Casilao, J. (2023, January 9). PNP: Theft, rape, physical injury most prevalent crimes in last 6 months. GMA News Online. https://www.gmanetwork.com/
• Chawla, N. V., Bowyer, K. W., Hall, L. O., & Kegelmeyer, W. P. (2002). SMOTE: Synthetic Minority Over-sampling Technique. Journal of Artificial Intelligence Research, 16, 321–357. https://doi.org/10.1613/jair.953
• Chen, T., & Guestrin, C. (2016). XGBoost: A Scalable Tree Boosting System. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 785–794. https://doi.org/10.1145/2939672.2939785
• Department of the Interior and Local Government. (n.d.). Angeles City Profile, accessed 28 October 2025, https://region3.dilg.gov.ph/
• Faeldon, E. (2019). philippines-json-maps [GeoJSON dataset]. GitHub. https://github.com/
• Gom-gom-o, J. (2024). CHALLENGES ENCOUNTERED BY THE POLICE OFFICERS IN CRIME PREVENTION IN TABUK CITY, KALINGA. EPRA International Journal of Multidisciplinary Research (IJMR), 10(5), Article 5.
• Jenga, K., Catal, C., & Kar, G. (2023). Machine learning in crime prediction. Journal of Ambient Intelligence and Humanized Computing, 14, 2887-2913. https://doi.org/10.1007/s12652-023-04530-y
• Philippine Statistics Authority. (2024). 2024 Census of Population (POPCEN): Population Counts Declared Official by the President, accessed 28 October 2025, https://psa.gov.ph/
• Safat, W., Asghar, S., & Gilani, S. (2021). Empirical Analysis for Crime Prediction and Forecasting Using Machine Learning and Deep Learning Techniques. IEEE Access, 1-1. https://doi.org/10.1109/ACCESS.2021.3078117
• Wubineh, B. Z. (2024). Crime analysis and prediction using machine-learning approach in the case of Hossana Police Commission. Security Journal, 37(4), 1269–1284. https://doi.org/10.1057/s41284-024-00416-6
• Yadhunath, R., Srikanth, S., Sudheer, A., & Palaniswamy, S. (2019). Identification of Criminal Activity Hotspots using Machine Learning to Aid in Effective Utilization of Police Patrolling in Cities with High Crime Rates. 2019 4th International Conference on Computational Systems and Information Technology for Sustainable Solution (CSITSS), 1–6. https://doi.org/10.1109/CSITSS47250.2019.9031057
• Yao, S., Wei, M., Yan, L., Wang, C., Dong, X., Liu, F., & Xiong, Y. (2020). Prediction of crime hotspots based on spatial factors of Random Forest. 2020 15th International Conference on Computer Science & Education (ICCSE), 811-815. https://doi.org/10.1109/ICCSE49874.2020.9201899
• Yunus, A., & Loo, J. (2024). London street crime analysis and prediction using crowdsourced dataset. Journal of Computational Mathematics and Data Science, 10, 100089. https://doi.org/10.1016/j.jcmds.2023.10008