论文部分内容阅读
检验一个随机变量是否服从正态分布有很多方法。例如,X~2检验法、Dn检验法以及被国家标准化委员会选为国家标准方法的W检验、D检验、偏度一峰度检验法等。这些方法的计算量都比较大。概率图纸法(亦被选为国家标准方法)直观而又简便,然而其精度不高,不能控制犯错误的概率。本文提出一个“差数”检验法,用来检验分布的对称性。该方法没有充分地利用样本信息,精度不高,但其计算量小,简单易行,还能控制犯第一类错误的概率。因此,它可以同概率图纸法一样作为精确检验前的预备检验。
There are many ways to test whether a random variable follows a normal distribution. For example, X ~ 2 test method, Dn test method and by the National Standardization Committee selected as the national standard W test, D test, skewness and kurtosis test. These methods are computationally intensive. Probability drawings (also selected as national standard methods) are intuitive and easy, yet their accuracy is not high enough to control the probability of making mistakes. This paper presents a “difference” test, used to test the distribution of symmetry. The method does not make full use of the sample information, the accuracy is not high, but its calculation is small, simple and easy, and can control the probability of making the first kind of error. As a result, it can be used as a pretest for accurate tests, just as the probability chart method.