正态概率纸检验的改进及推广  

Improvement and Promotion of Normal Probability Paper Test

作　　者：王丙参[1] 魏艳华[1] 丁恒飞[1] 

机构地区：[1]天水师范学院数学与统计学院,甘肃天水741001

出　　处：《统计与信息论坛》2018年第3期26-30,共5页Journal of Statistics and Information

基　　金：国家自然科学研究基金<反常扩散方程的高阶数值算法及其在天水地下水质研究中的应用>(11561060);甘肃省自然科学基金资助<黄河兰州段泥沙输运的分数阶动力学方程建模及其高阶快速算法研究>(17JR5RE009)

摘　　要：利用蒙特卡洛方法在正态概率纸检验中引入相关系数等指标,将定性判断与定量判断相结合,使得检验结果更加客观。最后,将正态概率纸检验及其改进推广到对数正态分布、指数分布、威布尔分布。概率纸检验的正确率与数据总体有关,数据总体与检验分布有差异且差异越小正确率越低。The Monte Carlo method is used to introduce the correlation coefficient and other indexes into the normal probability paper test,and the qualitative and quantitative judgment are combined to make the test results more objective.Finally,the normal probability paper test and its improvement are generalized to the logarithmic normal distribution,exponential distribution and Weibull distribution.The correct rate of probability paper test and the overall data are closely related.There is a difference between the overall data and the test distribution,the smaller the difference,the lower the correct rate.

关 键 词：正态分布 概率纸检验 蒙特卡洛 相关系数 威布尔分布 

分 类 号：O212[理学—概率论与数理统计]