基于联系数贴近度的区间型组合预测模型及其有效性

摘要

利用区间数和二元联系数的相互转化关系,把区间数组合预测问题转换成二元联系数组合预测问题.在联系数贴近度的最优准则下,建立基于联系数贴近度的区间型组合预测模型,分析了该模型的有效性理论,包括:基于联系数贴近度的区间型组合预测模型是非劣性组合预测、优性组合预测的充分条件定理,基于联系数贴近度的区间型组合预测模型的冗余预测方法的存在性和冗余方法的判定定理.对某省社会保障水平适度区间值进行组合预测的实证分析,结果显示所建立的模型能有效提高预测的精度.%By using the mutual transformation connection between interval number and dual number, the forecast problem of interval number combination is converted into the forecast problem of binary contact number combination.Under optimal criterion of the contact number, then founding an Interval combination forecast model based on Contact number close to the degree, and analyzing the effective theory of the model, including: Based on connection number type of interval combination forecast model yes-no a recessive combination forecast, Sufficient conditions theorem of optimum combination forecast, The judgment which is based on connection number type of interval combination forecasting model of the existence of redundant forecasting method and redundant method.Finally, the province social security level is moderate interval value combination forecast of the empirical analysis, and the result shows the established model is able to effectively improve the precision of the forecast.