Traffic crash prediction in Kano, Nigeria: multivariate long short-term memory method

Abstract

Accurate traffic crash prediction is crucial for implementing effective road safety measures. In this study, the performance of long short-term memory (LSTM) and multivariate LSTM (MLSTM) models in forecasting total crash count data in Kano state, Nigeria were compared. Human and vehicle factors, including speed violations, tyre bursts, brake failures, sign/light violations and phone use while driving, were incorporated as covariates in the MLSTM model. An autoregressive integrated moving average with exogenous variables (Arimax) model was used to investigate the effects of the covariates. The MLSTM model outperformed both the basic LSTM model and individual covariate models, emphasising the synergistic effect of considering a broad range of factors. The Arimax model results revealed that speed violation is significantly positively correlated with total crashes; the other covariates showed positive correlations but did not reach the statistical significance. The findings underscore the importance of a multivariate approach in enhancing traffic crash prediction. The MLSTM model's superior performance highlights the value of considering a comprehensive range of factors that influence crash occurrence to achieve more accurate predictions. Practical applications of these models could involve leveraging them for proactive traffic safety measures, including increased enforcement of traffic rules, improvements to road infrastructure and targeted driver education and campaigns.