Using the Method Combining PCA with BP Neural Network to Predict Water Demand for Urban Development

Combining Principal Component Analysis (PCA) with BP Neural Network, the paper has established a model to predict water demand for urban development with a demonstration in Hefei city. The results indicate that the error absolute value of prediction model is less than 0.9 percent with an ideal effect. Viewed from PCA results, the principal factors affecting urban water demand can be summarized up as economic development (first principal component F₁) and population size (second principal component F₂). By model training of BP network with the scores of F₁ and F₂ as inputs and water demand as outputs, we has provided three prediction programs, while we think the medium program is relatively better suitable for guiding Hefei´s water resources planning according to a comparative analysis on the balance between water supply and demand.