Chapter 32. Discriminant Analysis

Table of Contents

Introduction
Method description
The discriminant analysis model
Model assumptions
Usage
Data requirements
Model building and testing
Model application
Example
References

Introduction

Discriminant analysis is a purely statistical approach to the classification problem. The core function of the Discriminant Analysis Module is to support binary classification by the estimation of a so-called discriminant function. In the present settings the module supports binary linear discriminant analysis, so it can be used to perform binary classification tasks, similarly to the Logistic Regression module. However, the user should keep in mind that these two models are based on different assumptions concerning the learning data and thus their relative performance can vary, depending on whether these assumptions are fulfilled or not (refer also to the section model assumptions below).