论文部分内容阅读
在科学实验中,人们总希望耗费较少的人力、物力和财力,并在较短的时间内取得最优的结果。为此目的,曾提出了各种各样的试验设计,究竟哪种试验设计好呢?正交设计能适当减少试验的次数,并使统计简化;旋转设计保证了因子空间中同一球面上各点预测值的方差相等,排除了一部分误差的干扰。而D-最优设计,从统计意义上比较,是最好的试验计划。本文将简单介绍试验计划的评价和D-最优设计
In scientific experiments, people always hope to spend less manpower, material and financial resources, and in the shortest possible time to obtain the best results. To this end, a variety of experimental designs have been proposed. Which design is good? Orthogonal design can reduce the number of tests and simplify the statistics. Rotation design ensures that the points on the same sphere in the factor space The variance of the predicted values is equal, excluding part of the error interference. The D-optimal design, from a statistical point of comparison, is the best test plan. This article will briefly introduce the evaluation of the pilot program and the D-optimal design