Accurate traffic congestion forecasting is an indispensable element of urban transport systems. This paper suggests a machine learning model to predict rush-hour traffic congestion using a newly defined Traffic Congestion Index (M_TCI), incorporating traffic density as a crucial factor for congestion prediction. This study uses XGBoost algorithm with spatio-temporal and contextual features such as holidays and seasonality to enhance the model’s accuracy. The model focuses on long-term prediction, incorporating the day of the week, time, holiday and seasonality to predict daily road network performance. Results show that the model outperforms ensemble models- CatBoost, Gradient Boosting Machine (GBM) and LightGBM and achieves an accuracy of 90%. XGBoost performs better in handling large and high-dimensional datasets, making it a valuable tool for predicting traffic congestion and optimizing urban road networks.
Improved Road Traffic Congestion Prediction Using Machine Learning through Modified Index
Advances in intell. Systems Research
International Conference on Recent Advancements and Modernisations in Sustainable Intelligent Technologies and Applications ; 2025 ; Indore, India February 07, 2025 - February 08, 2025
Proceedings of the International Conference on Recent Advancement and Modernization in Sustainable Intelligent Technologies & Applications (RAMSITA-2025) ; Chapter : 13 ; 150-158
2025-05-25
9 pages
Article/Chapter (Book)
Electronic Resource
English
European Patent Office | 2025
|