Enhancing Crime Forecasting with Hybrid Approaches: Integrating Machine Learning and Statistical Models

• Paulo Cauas UFPE
• Ruam Pastor UFPE
• Paulo S. G. de Mattos Neto UFPE
• Filipe C. de L. Duarte UFPB

DOI: https://doi.org/10.5753/eniac.2025.12415

Abstract

Crime forecasting represents a significant challenge for public security. This article investigates the use of hybrid systems for crime time series forecasting, utilizing data from different Brazilian states. The proposed approach consists of combining Machine Learning (ML) models to capture nonlinear patterns in the time series, followed by the application of the statistical Autoregressive Integrated Moving Average (ARIMA) model to the forecast residuals, aiming to model remaining linear patterns. For evaluation, the proposed hybrid versions employed the following ML models: Long Short-Term Memory (LSTM), Multilayer Perceptron (MLP), and Support Vector Regression (SVR). The proposed versions (LSTM+ARIMA, MLP+ARIMA, and SVR+ARIMA) were compared with individual models and traditional hybrid approaches from the literature. The results demonstrate that the proposed hybrid approaches achieve superior performance, especially in scenarios with high complexity and a predominance of nonlinear patterns, reinforcing the importance of strategies that integrate different modeling paradigms. This study advances the field of crime forecasting by demonstrating the effectiveness of the proposed hybrid approach. The findings highlight the potential of hybrid systems as a robust and innovative foundation for decision support systems in public security, showing that the integration of different modeling paradigms can offer substantial gains in both accuracy and practical utility.

References

Abdulraheem, M., Awotunde, J., Oladipo, I., Adeleke, M., Ndunagu, J., Ayantola, J., and Mohammed, A. (2022). Crime rate prediction using the random forest algorithm. LAUTECH Journal of Engineering and Technology, 16(2):166–179.

Alwee, R., Hj Shamsuddin, S. M., and Sallehuddin, R. (2013). Hybrid support vector regression and autoregressive integrated moving average models improved by particle swarm optimization for property crime rates forecasting with economic indicators. The Scientific World Journal, 2013(1):951475.

Chainey, S. and Tompson, L. (2008). Crime mapping case studies: practice and research. John Wiley & Sons.

Dakalbab, F., Talib, M. A., Waraga, O. A., Nassif, A. B., Abbas, S., and Nasir, Q. (2022). Artificial intelligence & crime prediction: A systematic literature review. Social Sciences & Humanities Open, 6(1):100342.

Dave, E., Leonardo, A., Jeanice, M., and Hanafiah, N. (2021). Forecasting indonesia exports using a hybrid model arima-lstm. Procedia Computer Science, 179:480–487.

de Oliveira, J. F. L., Silva, E. G., and de Mattos Neto, P. S. G. (2021). A hybrid system based on dynamic selection for time series forecasting. IEEE Transactions on Neural Networks and Learning Systems, 33(8):3251–3263.

Duarte, F. C. d. L., de Mattos Neto, P. S., and Firmino, P. R. A. (2024). A hybrid recursive direct system for multi-step mortality rate forecasting. The Journal of Supercomputing, pages 1–34.

Gorr, W. and Harries, R. (2003). Introduction to crime forecasting. International Journal of Forecasting, 19(4):551–555.

Hajirahimi, Z. and Khashei, M. (2022). Sequence in hybridization of statistical and intelligent models in time series forecasting. Neural Processing Letters, 54(5):3619–3639.

Hodson, T. O. (2022). Root mean square error (rmse) or mean absolute error (mae): When to use them or not. Geoscientific Model Development Discussions, 2022:1–10.

İlgün, Esen Gül and Dener, M. (2025). Exploratory data analysis, time series analysis, crime type prediction, and trend forecasting in crime data using machine learning, deep learning, and statistical methods. Neural Computing and Applications. Published online: 20 March 2025.

Izidio, D. M., de Mattos Neto, P. S., Barbosa, L., de Oliveira, J. F., Marinho, M. H. d. N., and Rissi, G. F. (2021). Evolutionary hybrid system for energy consumption forecasting for smart meters. Energies, 14(7):1794.

Kang, H. W. and Kang, H. B. (2017). Prediction of crime occurrence from multi-modal data using deep learning. PLoS ONE, 12(4):e0176244.

Khairuddin, A. R., Alwee, R., and Haron, H. (2020). A comparative analysis of artificial intelligence techniques in forecasting violent crime rate. In IOP Conference Series: Materials Science and Engineering, volume 864, page 012056. IOP Publishing.

Khashei, M. and Hajirahimi, Z. (2019). A comparative study of series arima/mlp hybrid models for stock price forecasting. Communications in Statistics-Simulation and Computation, 48(9):2625–2640.

Lab, S. P. and Hirschel, J. D. (1988). Climatological conditions and crime: The forecast is. . . ? Justice Quarterly, 5(2):281–299.

Liu, M.-D., Ding, L., and Bai, Y.-L. (2021). Application of hybrid model based on empirical mode decomposition, novel recurrent neural networks and the arima to wind speed prediction. Energy Conversion and Management, 233:113917.

Mandalapu, V., Elluri, L., Vyas, P., and Roy, N. (2023). Crime prediction using machine learning and deep learning: A systematic review and future directions. IEEE Access, 11:60153–60170.

Manowska, A., Rybak, A., Dylong, A., and Pielot, J. (2021). Forecasting of natural gas consumption in poland based on arima-lstm hybrid model. Energies, 14(24):8597.

Mao, L., Du, W., Wen, S., Li, Q., Zhang, T., and Zhong, W. (2025). Crime forecasting: A spatio-temporal analysis with deep learning models. Journal of Computational Methods in Sciences and Engineering, page 14727978251337993.

Marzan, C. S., Baculo, M. J. C., de Dios Bulos, R., and Ruiz Jr, C. (2017). Time series analysis and crime pattern forecasting of city crime data. In Proceedings of the 1st International Conference on Algorithms, Computing and Systems, pages 113–118.

McDowall, D., Loftin, C., and Pate, M. (2012). Seasonal cycles in crime, and their variability. Journal of quantitative criminology, 28:389–410.

Muthamizharasan, M. and Ponnusamy, R. (2024). A comparative study of crime event forecasting using arima versus lstm model. J. Theor. Appl. Inf. Technol, 102:2162–2171.

Noor, T. H., Almars, A. M., Alwateer, M., Almaliki, M., Gad, I., and Atlam, E.-S. (2022). Sarima: a seasonal autoregressive integrated moving average model for crime analysis in saudi arabia. Electronics, 11(23):3986.

Olligschlaeger, A. M. (1997). Artificial neural networks and crime mapping. Crime mapping and crime prevention, 1:313.

Pai, P.-F. and Lin, C.-S. (2005). A hybrid arima and support vector machines model in stock price forecasting. Omega, 33(6):497–505.

Pastor, R. E., de Sales, J. P., Adriano Filho, F., and de Mattos Neto, P. S. (2023). Improving crime prediction through ensembles. In Encontro Nacional de Inteligência Artificial e Computacional (ENIAC), pages 1063–1073. SBC.

Resende, J. P. d. and Andrade, M. V. (2011). Crime social, castigo social: desigualdade de renda e taxas de criminalidade nos grandes municípios brasileiros. Estudos Econômicos (São Paulo), 41:173–195.

Rubio, L. and Alba, K. (2022). Forecasting selected colombian shares using a hybrid arima-svr model. Mathematics, 10(13):2181.

Safat, W., Asghar, S., and Gillani, S. A. (2021). Empirical analysis for crime prediction and forecasting using machine learning and deep learning techniques. IEEE access, 9:70080–70094.

Shah, N., Bhagat, N., and Shah, M. (2021). Crime forecasting: a machine learning and computer vision approach to crime prediction and prevention. Visual Computing for Industry, Biomedicine, and Art, 4(1):9.

Shi, J., Wang, S., Qu, P., and Shao, J. (2024). Time series prediction model using LSTM-Transformer neural network for mine water inflow. Scientific Reports, 14(1):18284.

Wang, W. and Lu, Y. (2018). Analysis of the mean absolute error (mae) and the root mean square error (rmse) in assessing rounding model. In IOP conference series: materials science and engineering, volume 324, page 012049. IOP Publishing.

Xu, Y., Fu, C., Kennedy, E., Jiang, S., and Owusu-Agyemang, S. (2018). The impact of street lights on spatial-temporal patterns of crime in detroit, michigan. Cities, 79:45–52.

Zhang, G. P. (2003). Time series forecasting using a hybrid ARIMA and neural network model. Neurocomputing, 50:159–175.

Zhang, J., Liu, H., Bai, W., and Li, X. (2024). A hybrid approach of wavelet transform, arima and lstm model for the share price index futures forecasting. The North American Journal of Economics and Finance, 69:102022.

Zhang, Y., Luo, L., Yang, J., Liu, D., Kong, R., and Feng, Y. (2019). A hybrid arima-svr approach for forecasting emergency patient flow. Journal of Ambient Intelligence and Humanized Computing, 10:3315–3323.