Optimal Screening Methods
从预测和决策理论角度设计了最优筛选方法，发现多元正态情形下最优筛选区域是筛选变量的二次函数，并用实例说明。
SUMMARY Optimal screening methods from a predictive point of view and a decision theoretical approach are devised and are shown to have a similar form. In the multivariate normal case the optimal screening region is found to be a quadratic function of the screening variables. The results are illustrated by an example.