A CRIME PREDICTION MODEL USING DEEP LEARNING BERT FOR LAW ENFORCEMENT AGENCIES IN NIGERIA

Read full text Summarise with VIRAT
Authors: RUQAYYAH ABUBAKAR HASSAN, Peter Buba Zirra
Cited by: 0
References: 4

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This study aims to develop a crime prediction model that combines Deep Learning, BERT, LSTM, and GCN to support proactive policing and enhance public safety. The proposed framework employs BERT to analyze contextual data in crime-related text, LSTM to capture temporal dynamics, and GCN to model relationships between crime types, locations, and entities. Datasets were collected from ngCERT (2019–2025), the Nigerian Police Force, and open-source reports from Vanguard and Premium Times. The system's performance was evaluated using accuracy, precision, recall, F1-score, and ROC metrics. The system achieved 97.3% accuracy, which is 10.3% higher than the base research models considered. The results indicate that the hybrid transformer-based model achieves higher performance than conventional Machine Learning methods, such as Random Forest, Support Vector Machine, K- Nearest Neighbors, and CART, for crime prediction. The findings highlight the capability of Deep Learning transformers to improve crime forecasting and offer a practical framework for enhancing resource allocation and decision-making within Nigerian law enforcement agencies. The model can also be integrated into a security information and event management platform to give real-time predictions.

Fields of Study & Subjects

Artificial Intelligence in Law Crime Patterns and Interventions Data Mining and Machine Learning Applications

Publication details

Journal / source
VUNOKLANG MULTIDISCIPLINARY JOURNAL OF SCIENCE AND TECHNOLOGY EDUCATION
Volume
14
Issue
1
Pages
127–138
Publication date
27 March 2026
Language
EN