判别分析与Logistic回归的模拟比较

An Analogue Comparison of Discriminant Analysis and Logistic Regression

利用随机模拟方法,研究判别分析和Logistic回归分类的回判正确率。模拟结果显示,Logistic回归的回判正确率优于判别分析。随着随机误差的增大,Logistic回归与判别分析的回判正确率差异逐渐减小。随机误差超过一定界限,Logistic回归的回判正确率低于判别分析。在随机模拟的基础上,引入修正Logistic回归分类,模拟结果显示,修正Logistic回归分类略优于Logistic回归。

By stochastic simulation,the discriminant accuracy rate of Discriminant analysis and Logistic regression is studied.The result shows that the discriminant accuracy rate of Logistic regression is better than Discriminant analysis.As random error becomes bigger,the differences between the discriminant accuracy rate of two methods gradually reduced.When random error exceeded a certain limit,the discriminant accuracy rate of Logistic regression was worse than Discriminant analysis.Furthermore,the amended Logistic regression is introduced,which showed the amended Logistic regression was slightly superior to Logistic regression.