基于L_p范数的加权几何平均组合预测

Weighted Geometric Average Combination Forecasting Method Based on the L_p Algorithm

组合预测模型的参数估计方法大都是在预测误差平方和最小准则之下建立的.针对误差平方和最小准则存在的不足,提出一种基于L_p范数的加权几何平均组合预测模型,并给出寻找最优L_p范数的蚁群算法.实例分析表明,该模型给出的组合预测方法为优性组合预测,从而证明了该模型的有效性.

At present,the combination forecast model parameter estimation methods are mostly based on minimizing the sum of squared errors criterion.In this paper,we first analyze the shortcomings of the error square and minimum standards,then establish the weighted geometric average combination model based on Lpnorm,and given the ant colony algorithm to find the optimal Lpnorm.analysis of the example shows that the combination forecast method of the model is given for optimal combination forecasting,to show the effectiveness of the proposed model.