Heterogeneous Data Processing for Bus Arrival Time Prediction

As the car number of Beijing break through five million, the mode share of public transportation sector is not large enough to ease congestion. To attract more people, the public transit should provide high quality service to the public. One way to achieve this is by developing intelligent transportation systems application to provide accurate bus arrival time prediction. Bus arrival time prediction has been a topic of research at home and abroad for years. However, little studies focus on the primary heterogeneous data processing which was the basic component of engineering practice. This study presents the method of primary data processing under heterogeneous conditions for bus arrival time prediction. The research also intends to test the proposed method using real-world bus data in Beijing, China, which performed effectively in terms of both prediction accuracy and computing time.