A Nonparametric Test Statistics for Two-Sample Problem

碩士論文[[abstract]]對於檢定兩組獨立樣本之分配函數的問題，目前已有多位學者研究過，較為熟悉的檢定統計量有： 1933年Kolmogorov提出一檢定統計量 來檢定一組樣本的分配函數，Kolmogorov提出兩組樣本的分配函數之檢定統計量。同樣檢定母體分配尺寸的統計量的有：1962年Klotz提出了一個用於檢定兩未知母體是否同為一分配的檢定統計量，2004年陳奕良利用兩樣本混合後，Y值落入X區間的觀念提出一個新的檢定統計量，來檢定分配的尺寸、以及是否為相同分配。 本論文主要是提出兩個檢定統計量，並以蒙地卡羅的方法模擬其檢定力。[[abstract]]The issue about testing on distribution function with two samples has been discussed by many scholars. However, the famous testing statistics in relation to this topic , suggested by Kolmogorov(1933). And the testing statistics based on scale are presented by Klotz(1962) . In 2005,Yi-Liang Chen suggest a new scale testing statistic,based on the number of the Y’s falls into intervals divided by X’s. In scale problem, our statistics compare the power by the Monte Carlo method in this study. In location problem, our statistics compare by the Monte Carlo method in this study