摘要
文章通过对组合预测模型权重求解方法的研究,给出了一种运用二次规划来求解组合预测模型权重的方法。该方法不仅克服了拉格朗日乘子法不能保证权重非负的不足,而且由于二次规划问题可借助MATLAB软件进行求解,从而简化了求解权重的计算过程。示例计算结果表明,组合预测模型的预测精度明显高于各单一预测模型,误差平方和小于各单一预测模型的误差平方和,从而验证了该方法的有效性。
In this paper, through the research on method of solving weight combination forecasting model, a method that used quadratic programming for weight of combination forecasting model was presented. The method not only overcame the shortage of the Lagrange multiplier method which couldn't ensure nonnegative condition of weight, but also simplified the calculating procedure of weight because quadratic programming problem could be solved by MATLAB software. The sampling calculation showed that the precision of combination forecasting model was higher than that of the single model, the error square sum of combination forecasting model was less than that of the single model, thereby, the validity of the method was proved.
出处
《东北农业大学学报》
CAS
CSCD
2008年第4期51-54,共4页
Journal of Northeast Agricultural University
关键词
组合预测
二次规划
负荷预测
权重
combination forecasting
quadratic programming
load forecasting
weight