TABLE OF CONTENTS Discriminant analysis is a powerful multivariate statistical method that can classify individual objects or units into groups. Specifically, in discriminant analysis, there are many individual objects and each object has several variables associated with it, that is, ( x 1, x 2, , x n). The whole population of objects can be partitioned into several groups with the values of variables strongly related to the group the object belongs to. Discriminant analysis is a method that classifies each object into groups depending on the values of the variables of each object. We give the following examples to illustrate this kind of a situation.
Example 6.1 : People s Weight Every person can be classified into one of the three groups underweight, normal weight, and overweight. In order to determine a person s weight category, we need to look into at least two variables, height ( x 1) and weight ( x 2). Table 6.1 gives a data set which has people s height and weight data in all the three weight categories.
Table 6.1: People s Weight Categories Weight category H (m)
x 1
W (kg)
x 2
Weight category H (m)
x 1
W (kg)
x 2
Weight category H (m)
x 1
W (kg)
x 2
N
1.52
48.5
O
1.53
61
U
1.54
45
N
1.61
59.6
O
1.60
64
U
1.63
47
N
1.65
65.3
O
1.66
82.6
U
1.67
47.4
N
1.71
64.3
O
