Optimizing Crime Prediction: Performance Comparison of 13 Machine Learning Models on Urban Crime Data
JOIV : International Journal on Informatics Visualization · 28.05.2026 · 10 (3) · 1101
DOI: 10.62527/joiv.10.3.4770 ↗Kurz gefasst
Fachartikel in der Zeitschrift JOIV : International Journal on Informatics Visualization, online erschienen am 28.05.2026. Verfasst von - Zulkarnaen und 7 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).
Abstract (Originalsprache)
Accurate crime prediction is critical for strengthening law enforcement and public safety measures in increasingly complicated urban settings. The purpose of this study is to assess a model's ability to reliably forecast crime categories using data from a dataset. The dataset provides a comprehensive view of criminal activity in a city by including information on each occurrence, such as date, time, location, victim demographics, and more. By applying machine learning to large datasets, we can anticipate crime categories, improve law enforcement strategies, and promote public safety initiatives. This study employs 13 machine learning algorithms to develop an accurate model, including SGD, K-Neighbors, BernoulliNB, DecisionTree, RandomForest, GradientBoosting, AdaBoost, Bagging, ExtraTrees, MLP, XGB, CatBoost, and LGBM. Each algorithm's performance was evaluated using six key metrics: accuracy, precision, recall, F1 Score, ROC AUC, and log loss. The study's findings indicate that the top-performing models are XGB, CatBoost, and LGBM. XGB has AUC, ACC, and Log Loss values of 0.9935, 0.954, and 0.13443. CatBoost's AUC, ACC, and Log Loss values are 0.9925, 0.94625, and 0.1597. The LGBM achieves AUC, ACC, and Log Loss of 0.9933, 0.9525, and 0.13998, respectively. This report provides practical recommendations to law enforcement agencies on how to select the optimal model for the efficient implementation of crime-mitigation techniques.
Zitieren
APA
Zulkarnaen, Nugroho, H., Wahyudi, S., Retnaningrum, D. H., Karno, A. S. B., Arman, S. A., Saputro, A. E., & Hastomo, W. (2026). Optimizing Crime Prediction: Performance Comparison of 13 Machine Learning Models on Urban Crime Data. JOIV : International Journal on Informatics Visualization, 10(3), 1101. https://doi.org/10.62527/joiv.10.3.4770
Vancouver
Zulkarnaen, Nugroho H, Wahyudi S, Retnaningrum DH, Karno ASB, Arman SA, et al. Optimizing Crime Prediction: Performance Comparison of 13 Machine Learning Models on Urban Crime Data. JOIV : International Journal on Informatics Visualization. 2026;10(3):1101. doi:10.62527/joiv.10.3.4770
BibTeX
@article{Zulkarnaen2026,
author = {Zulkarnaen, - and Nugroho, Hibnu and Wahyudi, Setya and Retnaningrum, Dwi Hapsari and Karno, Adhitio Satyo Bayangkari and Arman, Shevty Arbekti and Saputro, Ahmad Eko and Hastomo, Widi},
title = {Optimizing Crime Prediction: Performance Comparison of 13 Machine Learning Models on Urban Crime Data},
journal = {JOIV : International Journal on Informatics Visualization},
year = {2026},
volume = {10},
number = {3},
pages = {1101},
doi = {10.62527/joiv.10.3.4770}
}
EUVIMED