分析資料的統計方法有兩類：一種是無母數方法，另一種是有母數方法。雖然資料用無母數方法分析可以不假設任何特定的母體分布，但正確的使用有母數方法可以獲取較多的資訊。為能正確使用有母數方法，就必須根據資料建立分布的適合度檢定。本文分別在完整或右設限的一維度資料之下修正Kolmogorov和Cramer-von Mises統計式，在成對資料之下推廣修正Kolmogorov和Chi-square統計式進行資料分布的適合度檢定，此處的一維度資料考慮配適廣義伽瑪分布，成對資料則針對關聯結構函數做適合度檢定。本文以模擬的方法研究所提出適合度檢定的型I誤差率及檢定力的表現，最後以實例說明所提出檢定方法之應用。 There are two kinds of statistical methods for analyzing data: one is the nonparametric analysis and the other is the parametric analysis. We do not need to assume any particular form for the population distribution when we use a nonparametric method, however, correctly using a parametric method would produce more information on data analysis. To do so, we need to test the goodness-of-fit of a particular distribution based on the available data. In this paper, we construct goodness-of-fit tests for univariate and bivariate observations, respectively, with completely observed or right-censored data. Modifications of the Kolmogorov and Cramer-von Mises tests are proposed for testing the goodness-of-fit of the generalized gamma distribution for univariate data. Extensions of the Kolmogorov and Chi-square tests to testing the goodness-of-fit of a Copula function for bivariate data are then suggested. The results of a simulation study are presented for the investigation of type I error rates and powers of the proposed tests. Finally, the application of the tests is illustrated by using a real data set.
