集对分析聚类预测法及其应用

Method and application of set pair analysis classified prediction

提出一种新的预测方法—集对分析聚类预测法 .该方法融合了集对分析中的同异反模式识别的“择近原则”和聚类分析的基本思想进行分类预测 .文中将该方法应用于邮电业务总量预测的研究 ,考虑了邮电业务总量和第一、二、三产业的国内生产总值之间的关系 ,利用邮电业务总量和三个产业的生产总值的历史数据 ,建立了邮电业务量水平聚类预测的模型 ,从而得到邮电业务总量预测结果 .利用我国某地区的实际数据进行分析计算 ,并与其它预测方法比较 ,结果表明该预测方法是有效的 .

A new forecasting method, the SPA classified prediction, is proposed in this paper. The model is used to forecast the business total of posts and telecommunications. The history data are adopted to set up the connection degrees between the classified systems and the reference system. The result of the business total of posts and telecommunications forecast is obtained by applying the classified analysis methods. A set of real data is used, and the result is satisfied.