A Review of Traffic Flow Prediction Methods in Intelligent Transportation System Construction

With the ongoing development of Intelligent Transportation Systems (ITS), traffic flow prediction methods have become a cornerstone of its technology. Accurate traffic flow forecasting can significantly support traffic management and decision-making.However, the expanding demand for such systems presents new challenges for the development of traffic flow prediction technologies. This paper offers a comprehensive review of traffic flow prediction methods utilized in ITS, classifying them into three categories: statistical-based, machine learning-based, and deep learning-based methods. By analyzing the principles, advantages, limitations, and applications of each approach within ITS, and comparing them through practical case studies, the paper emphasizes the pivotal role of deep learning models in improving prediction accuracy, particularly in short-term traffic forecasting. Despite the notable success of deep learning methods in short-term predictions, challenges persist, such as model generalization across diverse traffic scenarios and difficulties in long-term traffic flow forecasting. The paper concludes by suggesting future research directions, aiming to provide new perspectives and enhance the development of ITS technologies.