Author(s): Huang Haiou; Zhang Wei
Linked Author(s):
Keywords: Waterway engineering; Forecasting; Artificial neural network
Abstract: The prediction on waterway transport volume is essential for the layout of water resource facilities and inland waterway infrastructures. The BPNN with momentum was applied to build a model for the prediction. The observed data between 1977 and 2004 in China were used to establish and test the prediction model both for short period and long term prediction. In the short period forecast, namely one year ahead forecast, the NN can give more accurate results comparing with the normally used models such as exponential smoothing, self-correlation, grey system and logistic time series regression. In long term prediction, it performed very well both in five year ahead and ten year ahead forecast, too. It can be concluded from this research that BPNN is a good way for the prediction of waterway transport volume.
Year: 2005