Comparison between some of linear classification models with practical application

Abstract

Linear discriminant analysis and logistic regression are the most widely used in multivariate statistical methods for analysis of data with categorical outcome variables .Both of them are appropriate for the development of linear  classification models .linear discriminant analysis has been that the data of explanatory variables must be distributed multivariate normal distribution. While logistic regression no assumptions on the distribution of the explanatory data. Hence ,It is assumed that logistic regression is the more flexible and more robust method in case of violations of these assumptions.

In this paper we have been focus for the comparison between three forms for classification data belongs two groups when the response variable with tow categorise only.

The first form is the linear discriminant function ,The second is the probability form which it is derivative as alternative for the linear discriminant function while the third form is the probability function model. Of the logistic regression the comparison between these methods is based on measure of the probability  of misclassification .We show that the results of the  probability form  of the logistic regression has minimum probability of misclassification through the application on the data of two types of (leukemia).