Intelligent Traffic Prediction by Combining Weather and Road Traffic Condition Information: A Deep Learning-Based Approach

An intelligent transportation system (ITS) is the collection and processing of traffic data that uses dynamic navigation to provide multi-mode urban dynamic traffic information. It helps drivers actively avoid congested sections, and rational use of truth resources as to achieve the purpose of time-saving, energy-saving, and environmental protection. In this paper, we use R studio platform processing models, such as Random Forest and Support Vector Machine to predict the traffic congestion rate and speed of the traffic flow. Among the traffic prediction models, in addition to considering the congestion of past traffic sections and road traffic conditions, the deciding factors of the prediction also considered weather type, date, average wind speed, and temperature. Different from the usual work, after adding more decision factors, the case study in Shenzhen shows that considering more influencing factors can significantly improve prediction accuracy. The simulation results also show that the proposed method is superior than the other methods in daily traffic flow prediction in terms of prediction accuracy.