棉花检测能力认证活动中统计模型的应用

Research on application of statistical model in cotton-inspection-proficiency testing

在以往棉花检测能力认证的统计分析中,一般都采用Z比分数,但是Z比分数在参考值的选取以及分析方面存在一定的局限性。针对这一问题,选取灰色聚类分析方法,立足于三角白化权函数,以棉花的上半部平均长度等5个聚类指标,参加能力认证活动的14个实验室为聚类对象,将聚类对象划分为稳定性较好和一般2类。经过在棉花HVI能力认证活动中的理论和实验验证,灰色聚类分析方法可以作为类似能力认证活动的一个统计模型。

In the past,we typically used Z-score in the statistics analysis of the cotton-inspection-proficiency testing.But Z-score has some limitations in the selection of the reference number and analysis.In response to the problem,this paper will apply the grey clustering analysis which depends on the trigonal whitening weight function to divide the cluster objects into better and general categories through taking the fourteen laboratories as the cluster objects,cotton′s upper half mean length,etc.as five cluster indicators.Through the theoretical and experimental verifying in the cotton HVI proficiency testing,it shows that the gray clustering analysis may be used as a statistical model for the similar proficiency testing in the future.