Journal of Engineering, Project, and Production Management, 2026, 16(6), 2025-258
Short-Term Traffic Flow Prediction Model for Smart Cities Based on Gradient Boosting Decision Trees
Lecturer, Department of
Traffic Management Engineering, Henan Police College, Zhengzhou, 450046,
China,
E-mail: zll1883893@163.com
Project Management
Received November 5, 2025; revised December 26, 2026; January 6, 2026; accepted June 21, 2026
Available online July 2, 2026
Abstract: Short-Term Traffic Flow (STTF) prediction is a pivotal technology in intelligent transportation management. This technology is crucial to the development of smart cities. However, traditional prediction models have shortcom-ings in modeling spatiotemporal dependencies, resisting overfitting, and adapting to external factors. Therefore, a smart city short-term traffic flow prediction model that fuses Bidirectional Long Short-Term Memory (Bi-LSTM) networks with gradient-boosted decision trees is proposed to improve prediction accuracy and robustness. The research utilizes Bi-LSTM networks to capture the bidirectional temporal characteristics of traffic flow, introduces an extreme gradient boosting algorithm framework, and combines radio frequency identification technology to achieve high-precision data collection. The experiment findings indicate that the model's forecasts closely align with the real-world data, with a maximum Mean Absolute Percentage Error (MAPE) of only 3.07%, far superior to traditional methods. The model can accurately capture the influence of weather factors: flow on rainy days is about 20% lower than on sunny days, and on rainstorm days, it is reduced by an additional 15%. After optimiza-tion, the model’s inference speed reaches 1-2 milliseconds per sample (on an NVIDIA V100 GPU, a high-performance data center GPU designed for AI and deep learning), which is more than 50% faster than the tradi-tional model. The introduced model exhibits remarkable precision and broad applicability for forecasting traffic flow in smart city environments, effectively addressing the lack of accuracy and reliability in current prediction methods under recurrent traffic patterns. It provides an efficient solution for short-term traffic flow prediction and demonstrates potential for future extension to abnormal event response and multi-city application verification.
Keywords: Bidirectional LSTM; gradient boosting decision tree; intelligent transportation systems; smart cities; traffic flow prediction.
Copyright © Journal of Engineering, Project, and Production Management (EPPM-Journal).
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.
Requests for reprints and permissions at eppm.journal@gmail.com.
Citation: Zhang, L. (2026). Short-Term Traffic Flow Prediction Model for Smart Cities Based on Gradient Boosting Decision Trees. Journal of Engineering, Project, and Production Management, 16(6), 2025-258.
