順序反應變數的累積羅吉斯迴歸分析

累積羅吉斯迴歸模型常用來推論順序型反應變數與風險因子之間的關係。針對研究中的反應變數為有順序性的情況下，本文分別彙整前瞻式累積羅吉斯迴歸模型的參數估計和適合度檢定，以及回溯式研究中累積羅吉斯迴歸模型的參數估計，並建構回溯式研究中模型的適合度檢定。為了檢測回溯式累積羅吉斯迴歸模型的合適性，本文將順序型反應變數視為兩類，然後應用Cheng and Chen (2004)的方法提出分數形式適合度檢定。經由模擬研究發現，即使在有限樣本下該檢定統計量亦表現良好。

關鍵字

前瞻式研究；回溯式研究；病例對照研究；適合度檢定；累積羅吉斯迴歸；最大概似估計量；順序型反應變數

並列摘要

In case-control studies, the cumulative logistic regression model is used popularly for inferring the relationship of ordinal response and risk factors. For the ordinal response, this paper summarizes the parameter estimation and goodness-of-fit of the cumulative logistic regression model in the prospective and retrospective studies, respectively. In order to test the goodness-of-fit of the retrospective cumulative logistic regression model, this paper divides the categories of response into two categories and then applies the method of Cheng and Chen (2004) to propose a score type goodness-of-fit test statistic. Simulation studies demonstrate that the proposed test performs well even in finite samples.

並列關鍵字

prospective study ； retrospective studies ； case-control studies ； goodness-of-fit test ； cumulative logistic regression ； maximum likelihood estimator ； ordinal response