Logistic 回归(七年制_余松林).pptVIP

Multivariate analysis Multiple models Linear regression Logistic regression Cox model Poisson regression Loglinear model Discriminant analysis ...... Choice of the tool according to the objectives, the study, and the variables 第十六章 Logistic 回归 Logistic Regression Lu Yunxia Dept. of Epidemiology and Biostatistics Public Health School Tongji Medical College Huazhong University of Science and Technology （一）53例接受手术的前列腺癌患者情况 （二）26例冠心病病人和28例对照进行病例?对照研究 26例冠心病病人和28例对照者进行病例?对照研究 Solution: Which methods could be used to analyze association between response variable and explanatory variables ? Review Linear regression Multiple linear regression Assumptions in linear regression Normal error terms （残差非正态） For a binary response variable, there are two cases: This clearly violates the normality assumption: two values only! Nonconstant error variance （方差不齐） Variances will differ at different levels of X Constraints on response function （因变量反映的是概率） Response function typically reflects probabilities 多元线性回归的适用条件 计量资料(均数)＝计量、等级、分类 Solution: Solution: Solution: 2. Logistic regression model When the response variable is binary, the shape of the response function is often sigmoidal Logistic response functions: 2. Logistic regression model An interesting and useful property of the logistic response function is that it can be linearized easily. If we make the logit transformation(Cox, 1970, logit变换) Then (with some algebra) 2. Logistic regression model P 发病的概率。 　　　为“机会”或“优势”(odds) logit P为优势之对数(log odds) Review Linear regression Maximum Likelihood Estimation (MLE)最大似然比法 MLE is a statistical method for estimating the coefficients of a model. The likelihood function (L) measures the probability of observing the particular set of dependent variable values (p1, p2, ..., pn) that occur in the sample:  L = Prob (p1* p2* * * pn) MLE involves finding the coefficients (?, ?) that makes the log of the likelihood function（似然