Optimization of GM (1, N) model based on numerical analysis and its application

The precision of GM (1, N) model depends on the building methods of background values on the extent. The paper, based on the approximate theory, adopts Newton-Cores Formula and Gauss-Legendre formula to optimize the background values of GM (1, N) model and the theory research indicates that the optimization can increase the precision of the GM (1, N) model. The optimized model is applied on the forecast of traffic noise. The research indicates that, compared to that of original model, the prediction precision of the optimized GM (1, N) model increases significantly. The average error deceases from original 5.749% to 4.738%. The analyzed theories and typical examples demonstrate the validity and applicability of the optimized method.