本文討論指數分布，柯西分布，t分布和常態分布之間的接近程度，我們以Pearson卡方及Kolmogorov-Sminov適合度檢定作為判斷的標準，基本概念如下:較不接近常態分布之分布在作常態性檢定時檢定力應較高。較不接近雙指數分布之分布在作雙指數性檢定時檢定力應較高。較不接近柯西分布之分布在作柯西性檢定時檢定力應較高。我們將用計算機模擬方式比較不同分布之檢定力。 In this paper , we discuss the closeness of two density functions by goodness of fit tests . The idea is as follows . Consider the test H:data is sampled from f . For those g that are close to f , the powers will be small . We will make comparisons between normal , double exponential , Cauchy and t distributions by simulations .
