Traffic flow forecasting based on rough set and neural network

As intelligent transportation systems (ITS) are implemented widely throughout the world, managers of transportation systems have access to large amounts of real-time status data. A variety of methods and techniques have been developed to forecast traffic flow. The traffic flow forecasting model based on neural network has been applied widely in ITS because of its high forecasting accuracy and self-learning ability. But the forecasting errors of the traffic flow curve near peaks are large, especially at the large slope difference on both side of the peak. As such, the traffic flow forecasting based on rough set and neural network is proposed. The traffic flow in the current time interval, traffic flow in the previous time interval, traffic flow deviation between the current time interval and the previous time interval and current time is regarded as an input of a neural network respectively. The forecasting traffic flow in the following time interval is the output of the neural network. The trained neural network is the traffic flow forecasting model based on neural network. Then, the forecasting traffic flow in the following time interval obtained by the neural network based traffic flow forecasting model is compensated by rough set to increase the forecasting accuracy. The simulation experiments show that the presented traffic flow forecasting based on rough set and neural network can improve the forecasting accuracy significantly.