D.W. Seng, J.W. Peng, J. Chen, N. Zheng
Arrival time forecasts can provide lots of help in improving the image of city. The public resident’s satisfaction with public transport system, and reduce the number of private cars. In the past, bus forecasts were based on a single static or dynamic prediction algorithm didn’t have real-time or stability. In this paper, Support Vector Machine algorithm (SVM) are used to predict the results through historical , and then use Kalman filter (KF) algorithm to predict the results of dynamic correction, while combine static and dynamic algorithms.
Keywords: bus station forecast; SVM algorithm; KF algorithm; combine static and dynamic algorithms