本文在优性组合预测方法存在判定理论的基础上,提出了求解非负权重最优组合预测问题的迭代算法。
Based on the concept of superior combination forecasting method and on the concept of redundant method, several problems about combined prediction of non-negative weight are studied in this peper.
结果表明:基于神经网络组合预测模型不仅比单一预测方法能够从整体上提高预测精度,而且能避免最优组合预测模型有时出现负权重的不足。
The result shows that this combined model can improve the precision of forecasting and avoid the shortage of negative weight in the process of optimal mix forecasting model.
应用推荐