Search

Urban traffic flow prediction has received much attention in the past few years, especially after the availability of huge traffic data. In addition, the efficacy of some existing traffic flow techniques heavily depends on some influential external factors (like weather, geographical information, point-of-interests (POIs), or road information). However, exploitable data containing such auxiliary information are extremely limited. One primary reason is that some rising cities lack the resources necessary for collecting such data. Therefore, it is difficult to pick an effective method that can accurately predict the traffic state when the data is limited. Due to that, this paper proposes a framework for achieving accurate predictions based on four data characteristics and the temporal correlation of neighbors without using additional influential factors. Specifically, we first propose a novel Multi-Region Correlation (MRC) method to make regions cooperate and share their traffic flow and patterns. Then, a deep recurrent network is used to capture the neighbors' temporal correlation by integrating their traffic history that flow among regions and their neighbors. When analyzing the traffic data history, we model four data characteristics (seasonality, trend, residual, and cyclic). And lastly, we employ a Multiple Linear Regression Unit (MLRU) to predict the future traffic directly from the neighbors' traffic. We conducted extensive experiments on two real-world datasets collected from two major cities in China, Chengdu and Xi'an. The results demonstrate that the proposed model can achieve superior performance compared to many existing models in terms of accuracy and efficiency. [ABSTRACT FROM AUTHOR]

As a flexible public transportation in urban areas, taxis play an important role in providing comfortable and convenient services for passengers. Due to the existence of the imbalance between supply of drivers and demand of passengers, an accurate fine-grained taxi demand prediction in real time can help guide drivers to plan their routes and reduce the waiting time of passengers. Recently, several methods based on deep neural networks have been provided to predict taxi demands. However, these works are limited in properly incorporating multi-view features of taxi demands together, with considering the influences of context information. In this paper, we propose a convolutional recurrent network model for fine-grained taxi demand prediction. Local convolutional layers and gated recurrent units are employed in our model to extract multi-view spatial–temporal features of taxi demands. Moreover, a novel context-aware attention module is designed to incorporate the predictions of each region with considering its contextual information, which is our first attempt. We also conduct comprehensive experiments based on multiple real-world datasets in New York City and Chengdu. The experimental results show that our model outperforms state-of-the-art methods, and validate the usefulness of each module in our model. [ABSTRACT FROM AUTHOR]