碩士 / 國立中興大學 / 應用數學系 / 93 / The method of regression is a kind of indispensable analysis tool, especially when we discuss the relation between a response variable and one or more explanatory variables. But traditional linear regression can only be impolyed when the response variable is continuous. If the response variable is discrete or categorical, then the logistic regression model can be used for solving this kind of problem.
This article discusses the ploytomous logistic regression models. The relationship between a response variable and explanatory variable, among them explain variable is a continuous type or the discrete type, and response variable is multiple discrete. We only consider two types of models for ordered response variable, the adjacent-category logistic model and continuation-ratio logistic model. Under each model, we use computer to simulate 1000 data sets. For each data set this two types of models are used to estimate the relevant parameters, and their standard errors. The percentage bias comes to assess and compare these two types of models. At the same time we come to assess and compare the suitability of these two types of models with empirical significance level and empirical power.
| Identifer | oai:union.ndltd.org:TW/093NCHU0507014 |
| Creators | liou tzy hsing, 劉姿杏 |
| Contributors | , 林良盈, 陳齊康 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | en_US |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 40 |
Page generated in 0.0178 seconds