An effective dynamic spatiotemporal framework with external features information for traffic prediction

Abstract

Traffic prediction is necessary for management departments to dispatch vehicles and for drivers to avoid congested roads. Many traffic forecasting methods based on deep learning have been proposed in recent years, and their main aim is to solve the problem of spatial dependencies and temporal dynamics. This paper proposes a useful dynamic model to predict the urban traffic volume by combining fully bidirectional LSTM, a complex attention mechanism, and external features, including weather conditions and events. First, we adopt bidirectional LSTM to obtain temporal dependencies of traffic volume dynamically in each layer, which is different from the hybrid methods combining bidirectional and unidirectional approaches. Second, we use a more elaborate attention mechanism to learn short-term and long-term periodic temporal dependencies. Finally, we collect weather condition and event information as external features to further improve the prediction precision. The experimental results show that the proposed model improves the prediction precision by approximately 3-7 percent on the NYC-Taxi and NYC-Bike datasets compared to the most recently developed method and is therefore a useful tool for urban traffic prediction.