Short Term Traffic Flow Prediction Using Deep Learning Approach

Intelligent transportation systems helps travellers reach their destination at an estimated time. Smart cities have deployed latest technologies to satisfy the needs of travellers through efficient navigation. Adverse weather conditions may lead to increase in traffic congestion as well as accidents. Rainfall, Snow and wet pavements are a major cause for the traffic flow during winter across all areas. Several methods have been researched at a deep level for predicting traffic flow and still an efficient method to satisfy real world traffic problems have not been identified. The accident and weather data are investigated in this paper to predict the flow of traffic in Leeds city. Three models were developed through deep learning neural network algorithm for predicting traffic flow. The predicted Deep neural network model is compared with the superior support vector machine model and the accuracy is higher for the Deep neural networks.