Short-Term Traffic Flow Prediction with Recurrent Mixture Density Network

Traffic situation awareness is the key factor for intelligent transportation systems (ITS) and smart city. Short-term traffic flow prediction is one of the challenging tasks of traffic situation awareness, which is useful for route planning, traffic congestion alleviation, emission reduction, and so on. Over the past few years, ubiquitous location acquisition techniques and sensors digitized the road networks and generated spatiotemporal data. Massive traffic data provide an opportunity for short-term traffic flow prediction in a data-driven manner. Most of the existing short-term traffic flow prediction methods can be divided into two categories: nonparametric and parametric. Traditional parametric methods failed to obtain accurate prediction, due to the nonlinear and stochastic characteristics of short-term traffic flow. Recently, deep learning methods have been studied widely in the fields of short-term prediction. These nonparametric methods yielded promising results in practical experiments. Motivated by the current study status, we dedicate this paper to a short-term traffic flow prediction approach based on the recurrent mixture density network, the combination of recurrent neural network (RNN), and mixture density network (MDN). This approach is implemented on real-world traffic flow data and demonstrates the prominent superiority. To the best of our knowledge, this is the first time that the recurrent mixture density network is applied to a real-world short-term traffic flow prediction task.