Data Fusion for Trip Prediction in Transit Information System

In the paper a transit information system for trip prediction is introduced. Several models, including ARIMA, neural network and data fusion, can be employed to forecast the number of passengers entering or leaving a station in a period in transit networks. The data fusion model combines two individual predictions from ARIMA and neural networks models. The objective of data fusion is to provide a better solution than could otherwise be achieved from the use of single-source data alone. Realistic transit networks in Guangzhou are selected as case studies in the development of the system.