针对电力负荷预测序列存在随机增长特性和非线性波动性,灰色神经网络模型能够有效地反映序列增长特性和对非线性关系进行拟合。引入灰色关联度作为确定组合预测权重的依据,以序列整体拟合优化为目标,从而找到能够代表序列内在变化规律性的组合权重,使总体预测精度得到提高。
As to the characteristics of random-increase and non-linear fluctuation of the load forecast sequence, the gray neural network model can effetively reflect the increase characteristic of the sequence and fit the non-linear relations. The gray correlation is introduced to determine the combined forecast weight, and fitting and optimization of the whole sequence is considered as the objective. The combined weight representing the inner changing rides of the sequence is consequently found out, and the forecast accuracy is improved.