Comparison of Traffic Flow Prediction Models Based on Deep Learning
Traffic flow prediction is a challenging task in Intelligent Transportation System (ITS). Accurate information of traffic flow help travelers to plan their routes wisely. It can help in reducing traffic congestion and improves efficiency. A number of traffic flow techniques exist but they fail to provide promising results that is because of their shallow learning architecture. When we compare these shallow architectures to deep learning architectures they are lack of feature learning capability. In this paper, we have introduced four deep learning architectures; LSTM,I-LSTM, RNN-LSTM and CNN-LSTM to predict the flow of traffic. These proposed models are applied on real-time traffic data collected from Jaipur, Rajasthan. It includes the comparison of all these architectures to find out the best one.