Logistic回归在判别分析中的新用法  

A New Application of Logistic Regression to Discriminant Analysis

作　　者：王国富[1] 张庆丰[1] 

机构地区：[1]安阳工学院,河南安阳455000

出　　处：《安阳工学院学报》2009年第6期89-91,共3页Journal of Anyang Institute of Technology

摘　　要：在非线性回归模型中,当响应变量为分类变量尤其是二分类变量时所采用的Logistic回归模型是一个很有应用前景的模型。它可以克服线性回归的一些缺点,提高判别分析的效率。通过分析Logistic回归应用于判别分析的特点提出采用二次判别法来提高判别效率,最后通过实际案例验证了所提做法的有效性。In the non-linear regression model, when the response variable is category variable, especially binary variables, logistic regression model is promising. It can overcome some of the shortcomings of linear regression and improve the efficiency of discriminant analysis. By analyzing the characteristics of the application of logistic Regression to discriminant analysis, the paper proposes that the quadratic discriminant can be adopted to improve the efficiency. In addition, the actual case is analyzed to verify the validity of the proposed approach.

关 键 词：LOGISTIC回归 判别分析:二次判别 

分 类 号：C913.9[经济管理]