Analysis of Criminal Patterns in Police Report Narratives using Spectral Clustering with K-means
Resumo
The high volume and heterogeneity of police report narratives in Brazil pose challenges for manual analysis and investigative prioritization. This work proposes an approach for identifying criminal patterns using clustering techniques applied to unstructured textual data. The methodology integrates Spectral Clustering with K-means, leveraging MPNet embeddings for vector representation and UMAP for dimensionality reduction. The resulting six clusters revealed thematic coherence, highlighting patterns such as bank fraud, judicial scams, social media crimes, and account hacking. Comparative experiments with Agglomerative Clustering were conducted using different linkage strategies, with Spectral Clustering achieving the highest silhouette score (0.38), indicating better-defined groups. A manual qualitative analysis of samples from each cluster supported the thematic distinctions. The study demonstrates that automatic clustering can contribute to investigative triage, offering relevant insights for public security applications.
Referências
Aggarwal, Charu C and Zhai, ChengXiang, A survey of text clustering algorithms, Springer, pp.77-128, 2012. DOI: 10.1007/978-1-4614-3223-4_4
Bellman, Richard E and Dreyfus, Stuart E, Applied dynamic programming, Princeton university press, 1957.
Chainey, Spencer, Crime mapping, Springer New York, Encyclopedia of Criminology and Criminal Justice, pp.699-709, 2013. DOI: 10.1007/978-1-4614-5690-2_317
Jain, Anil K, Data clustering: 50 years beyond K-means, Pattern recognition letters, Elsevier, pp.651-666, 2010. DOI: 10.1016/j.patrec.2009.09.011
Janez-Martino, Francisco and Alaiz-Rodriguez, Rocio and Gonzalez-Castro, Victor and Fidalgo, Eduardo and Alegre, Enrique, Classifying spam emails using agglomerative hierarchical clustering and a topic-based approach, Elsevier, Applied Soft Computing, 2023. DOI: 10.1016/j.asoc.2023.110226
Joshi, Anant and Sabitha, A Sai and Choudhury, Tanupriya, Crime analysis using K-means clustering, 2017 3rd International conference on computational intelligence and networks (CINE), pp.33-39, 2017. DOI: 10.1109/CINE.2017.23
Lal Beejal, Chaman and Ahmed, Awais and Siyal, Reshma and Kumar, Suresh and Aftab, Shagufta and Jamali, Arshad, Text Clustering using K-MEAN, International Journal of Advanced Trends in Computer Science and Engineering, pp. 2892-2897, 2021. DOI: 10.30534/ijatcse/2021/371042021
McInnes, Leland and Healy, John and Melville, James, Umap: Uniform manifold approximation and projection for dimension reduction, Journal of Open Source Software, pp. 861, 2025. DOI: 10.21105/joss.00861
Muennighoff, Niklas and Tazi, Nouamane and Magne, Loic and Reimers, Nils, Mteb: Massive text embedding benchmark, Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics, Dubrovnik, Croatia, 2023. DOI: 10.18653/v1/2023.eacl-main.148
Oliveira, Fabiano R and Zanusso, Maria B, Clusterização de ocorrências policiais utilizando k-means e um mapa auto-organizável, CBRN, 2005. DOI: 10.21528/CBRN2005-006
Reimers, N.; Gurevych, I., Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks, Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing, pp. 3982-3992, 2019. DOI: 10.18653/v1/D19-1410
Rousseeuw, Peter J., Silhouettes: a graphical aid to the interpretation and validation of cluster analysis, Elsevier, Journal of computational and applied mathematics, pp. 53-65, 1987. DOI: 10.1016/0377-0427(87)90125-7
SSPMT - Sistema de Registro de Ocorências Policiais do Estado de Mato Grosso. Dados extraídos do módulo SROP, referente ao registro de boletins de ocorrência. Cuiabá: SSP-MT, 2025. Dados obtidos via acesso interno, 2025.
Von Luxburg, Ulrike, A tutorial on spectral clustering, Statistics and computing, Springer pp. 395-416, 2007. DOI: 10.1007/s11222-007-9033-z
