摘要
通过等距采样和随机采样方式分别计算不同自由度下抽样分布与其渐近正态分布在分布函数上的绝对偏差,并建立绝对偏差平方和均值或平均绝对偏差与抽样分布自由度之间的非线性回归方程,结合各种直观有效的图形趋势分析,最终给出满足指定偏差要求时可用正态分布近似计算的最小自由度的估计.
The absolute deviations between sampling distribution with different degrees of freedom and asymptotic normal distribution were calculated based on equidistant sampling and random sampling. We established a nonlinear regression between the sum of the square of absolute deviation or the average of absolute deviation and degree of freedom of sampling distribution. Combining with the effective graphical trend analysis, we gave out the estimation of the minimum degree of freedom by using normal distribution approximation under certain condition or specified deviation requirements.
出处
《大学数学》
2017年第3期81-88,共8页
College Mathematics
基金
福建省本科高校教育教学改革研究项目(JAS151395)
国家自然科学基金青年基金(11301084)
福州大学第九批高等教育教学改革工程(52001024
52001069)
研究生优质课程建设项目(52004634
52004612)
关键词
随机模拟
抽样分布
自由度
非线性回归
stochastic simulation
sampling distribution
degree of freedom
nonlinear regression
