摘要
在能力验证活动中,针对稳健统计方法造成分类不精确的问题,利用核密度估计函数设计了一种纵向分类的评价标准.该方法以实验室测得的原始数据,运用核函数估计数据所服从的密度函数,然后计算出各测量值对应的函数值,通过函数值的大小达到分类的效果.实验结果表明:该方法增加了测量结果的满意度,弥补了稳健统计方法只能对正态分布数据进行分类的不足,扩大了数据分类的范围.
In proficiency testing activities, to the problems of inaccurate classification caused by robust statisti- cal methods, a vertical segment of the evaluation criteria is designed using kernel density estimation function. With the raw data measured in the laboratory, using kernel density function estimated the density function of the data obeyed, and then calculated the corresponding function value of the measured value, by the function value of the size, to achieve the effect of classification. The results show that this method increases the degree of satisfaction of the measurement results, makes up the lack of robust statistical method only to classify the normal data and expends the scope of data classification.
出处
《中北大学学报（自然科学版）》
CAS
北大核心
2013年第5期586-588,596,共4页
Journal of North University of China(Natural Science Edition)
基金
国家质量监督检验检疫总局资助项目(2013IK139)
关键词
核密度估计
正态检验
稳健统计
z比分数
kernel density estimation
normal test
robust statistics
z-score
