Road Traffic Accident Severity Prediction Using Machine Learning

This work focuses on analysing accident severity using machine learning techniques based on multiple influencing factors. The implementation of multiple machine learning models, including SVM, Random Forest, XGBoost, and Artificial Neural Network, allowed for a comparative analysis of their performance. Among these models, XGBoost emerged as the most effective algorithm, achieving the highest accuracy of 95%. This highlights the strength of boosting techniques in handling structured datasets and capturing complex relationships between features. The findings of this research emphasize that accident severity is influenced by multiple factors, including vehicle speed, weather conditions, visibility, and road conditions. By analysing these factors simultaneously, the proposed system provides a more reliable prediction compared to traditional statistical methods. One of the key contributions of this study is the development of a scalable and adaptable prediction framework that can be integrated into intelligent transportation systems. Such systems can assist traffic authorities in identifying high-risk scenarios and implementing preventive measures such as speed control, traffic regulation, and improved road infrastructure. Furthermore, this research highlights the importance of data-driven decision-making in road safety management. By leveraging machine learning techniques, policymakers can gain valuable insights into accident patterns and develop targeted strategies to reduce accident severity. Although the study uses a synthetic dataset, the model design ensures that it closely reflects real-world conditions. Future work can focus on integrating real-world datasets from Indian traffic authorities to further enhance model performance and reliability. In conclusion, the proposed machine learning-based approach provides an effective solution for accident severity prediction and has the potential to significantly improve road safety outcomes when implemented in real-world systems.

[1]  J. Zhang, Y. Wang, and L. Chen, “A New Framework of Vehicle Collision Prediction by Combining SVM and HMM,” IEEE Transactions on Intelligent Transportation Systems, vol. 21, no. 5, pp. 1234–1245, 2020. [2] S. Kumar and R. Singh, “An Automated Approach for Predicting Road Traffic Accident Severity Using Transformer Learning and Explainable AI,” IEEE Access, vol. 11, pp. 45678–45689, 2023. [3] M. Patel, A. Shah, and K. Mehta, “Automatic Accident Detection, Segmentation and Duration Prediction Using Machine Learning,” in Proc. IEEE Int. Conf. Intelligent Systems, 2021, pp. 210–215. [4]  Y. Li, X. Zhao, and H. Liu, “Interpretable Traffic Accident Prediction Using Attention-Based Spatial-Temporal Multi-Graph Learning,” IEEE Transactions on Neural Networks and Learning Systems, vol. 33, no. 8, pp. 3456–3468, 2022. [5]  A. Sharma, P. Verma, and S. Gupta, “Predicting the Type of Road Traffic Accident for Test Scenario Generation,” in Proc. IEEE Int. Conf. Smart Transportation, 2021, pp. 98–103. [6]  H. Wang, J. Liu, and Z. Sun, “RFCNN Traffic Accident Severity Prediction Based on Decision Level Fusion of Machine and Deep Learning Model,” IEEE Access, vol. 10, pp. 22345–22356, 2022. [7]  K. Verma, R. Tiwari, and M. Singh, “Road Traffic Accident Risk Prediction and Key Factor Identification Framework Based on Explainable Deep Learning,” IEEE Access, vol. 11, pp. 33456–33467, 2023. [8]  L. Chen, F. Wu, and Y. Zhang, “Traffic Accident Risk Prediction of Tunnel Based on Multi-Source Heterogeneous Data Fusion,” IEEE Transactions on Intelligent Transportation Systems, vol. 23, no. 4, pp. 2789–2798, 2022. [9]  R. Gupta, S. Das, and P. Roy, “Traffic Accident Risk Prediction via Multi-View Multi-Task Spatio-Temporal Networks,” IEEE Transactions on Intelligent Transportation Systems, vol. 22, no. 7, pp. 4123–4135, 2021. [10] P. Singh and A. Mishra, “Road Accident Prediction Using Machine Learning Techniques,” International Journal of Engineering Research, vol. 9, no. 3, pp. 145–150, 2020.