组合预测中基于粗糙集理论的权值确定方法

权系数的确定问题是组合预测方法中的关键和难点，提出了一种基于粗糙集理论的权系数确定方法，将权系数确定问题转化为粗糙集中属性重要性评价问题，建立了关于组合预测方法的关系数据模型，通过属性值特征化建立了知识系统，在数据驱动下通过分析预测方法对预测对象的依赖性和重要性，计算出组合预测模型的权系数，该方法克服了传统权系数确定方法的主观性，避免了线性或非线性极值问题的数值计算，使得组合预测方法具客观性，所给案例说明了所提方法 的有效性。

Determination to Weighting Coefficient of Combination Forecast Based on Rough Set Theory

The problem of determination to weighting coefficient is a key and difficulty for combination forecast. A method of determining weighting coefficient based on rough set theory is showed in this paper. Determining weighting coefficient is translated into estimating significance of attributes among rough set. A relation data model about combination forecast is established. Knowledge systems are built through making attribute value into eigenvalue. Under data moving, the weighting coefficients of a combination forecast model are computed by analyzing the dependence and significance of forecasting method for the predicted object. The proposed approach overcomes the subjectivity of traditional determination to weighting coefficient, avoids computing linear or nonlinear extremum problem and makes combination forecast more objective. The validity of the proposed approach is verified with a case.