摘要
在农业多因素试验中，应用二次通用旋转组合设计与频数分析方法寻找最佳农艺措施应注意两点：①如果回归剩余标准差远远小于约束条件的长度，那么统计区域必须选在单位球体内部（包括球面），或者选在外部的一个确定“球面”上。②如果回归剩余标准差与约束条件长度相差不大，则应以实际产量的置信区间和约束条件相比较进行频数统计，这样才能保证统计结论有较高的可靠性。
In experiments of the agricultural multi-factors ,the methods of Quadratic Combinatorial Designs with Revolvement and in Common use and Frequency Analysis are applied to seek the best agriculturalmeasures.The following points must be paid atlention to. 1. If the regressive residual standard deviation is far smaller than the length of the restritiveconditions , the statistical region has to be chosen inside theunitspheroid (including the sphere)or onan outside definite’ sphere’.In many applied articles , choosing the statistical region within thesupercube’ from-r to+r is not suitdble to the theoretical requirements.2. If there’s hardly any difference betweenthe regressive residual standard deviation and thelength of the restrictive conditions , the incredible region of actual yield is compared with the restrictiveconditions to make frequency statistics. Therefore the way can ensure the statistical conclusion withthe higher reliability.
出处
《西北农业学报》
CSCD
1994年第4期70-73,共4页
Acta Agriculturae Boreali-occidentalis Sinica
关键词
农业
农艺
频数分析
旋转组合
设计
Requency analysis
Statistic region
reliability of statistical conclusion
