基于灰色聚类的组合预测单项模型选择

建立组合预测模型关键是单项预测方法的筛选,本文将采用灰色聚类分析方法对单项预测模型组进行筛选.首先依据所研究的实际问题,建立多个单项预测模型,然后计算每一个单项预测方法的点拟合相对误差,最后再利用这些数据,借助灰色聚类方法实行对比评估,构建组合预测模型需求的各个单项模型,以增强该模型预测的准确性.应用实例的分析表明该方法是可行的和有效的.

The Selection of Single Model in Combination Forecasting Based on Grey Clustering

The key of establishing a combined forecasting model is the selection of single predicted method. In this paper,grey clustering analysis is applied to the selection of single forecasting model group. Reasonable single forecasting models are firstly established in accordance with the practical problems. Then the fitting relative errors from each single forecasting model are calculated. The data from the calculations are compared with each other by using grey clustering method to construct the individual model needed in each combination forecasting model. This practice will improve the predictive validity of combination forecasting. The analysis of application examples shows that the method is feasible and effective.