基于相对熵的最优区间型组合预测模型

Optimal Interval Combination Forecasting Model Based on Relative Entropy

文章将传统的区间数转化为二元联系数,分别以二元联系数的同部序列和异部序列作为研究对象,选取相对熵作为优化准则,引入偏好系数,建立基于相对熵的最优区间型组合预测模型,并给出该模型是优性组合预测的性质定理。实例分析结果显示,所构建的区间型组合预测模型可以显著地减少预测误差,有效提高预测精度。

This paper transforms the traditional interval numbers into binary relation numbers and takes the same part sequence and different part sequence of binary relation numbers as the research objects respectively, with the relative entropy chosen as the optimization criterion, introducing preference coefficient to establish the optimal interval combination prediction model based on relative entropy. Finally, the paper presents the property theorem of optimal combination prediction. Example analysis results show that the interval combination prediction model can significantly reduce the prediction error and improve the prediction accuracy.