论文部分内容阅读
首先,针对现有GMDH算法的不足做了两点改进:一是利用初始变量添加法来防止有用信息丢失、改善模型结构、提高模型精度,二是用权系数向量法或符号表达式法直接给出GMDH输出的显式表达式;然后,将改进的GMDH算法应用于煤炭价格影响因素的实证研究。结果显示改进的GMDH算法能够更真实地反映实际的煤炭价格模型结构,不仅有更高的拟合精度,而且有良好的外推能力,能够准确预测煤炭价格的变动趋势。
First of all, two improvements have been made to the deficiency of the existing GMDH algorithm: one is to use the initial variable adding method to prevent loss of useful information, improve the model structure and improve the accuracy of the model; secondly, use weight vector method or symbolic expression method directly The explicit expression of GMDH output is obtained. Then, the improved GMDH algorithm is applied to the empirical study of the factors affecting the coal price. The results show that the improved GMDH algorithm can more truly reflect the actual coal price model structure, which not only has a higher fitting accuracy but also has a good extrapolation ability and can accurately predict the trend of coal price changes.