Sunteți pe pagina 1din 8

DISCRIMINANT

ANALYSIS

POOJA BADERIA
MBA (BE)
ROLL NOS: 767

OBJECTIVE :
To understand smokers and non-smokers class differences and to predict the likelihood that
a particular entity will belong to a particular class or group based on blood pressure and
cholestrol.

Purpose:
1) The main purpose is to classify a person into smokers and non-smokers class.

2) A second purpose of the discriminant analysis is to study the relationship between


smokers and non-smokers class on the chloestrol and blood pressure variables which is used
to predict the group membership.

DATA:
I have used a secondary data , with 49 observations on three variables. The dependent
variables are blood pressure and chloestrol . The categorical variable is a person with two
different consumption habits 1)smoker, 2) non-smoker

The dummy variable for the smoker is 1 , otherwise

Research methodology:
Descriptive statistics
correlation
Discriminant analysis

Analysis has been done by the statistical tool STATA.

DATA
Cholestr Blood
Smoking ol Pressure
1 193 126
0 210 120
0 196 128
0 208 129
1 188 119
0 206 136
1 240 131
1 215 163
0 201 132
1 215 138
1 199 148
1 210 115
1 242 149
0 193 142
1 208 156
1 253 135
0 288 156
1 164 153
0 214 122
1 220 142
1 194 122
0 205 119
0 233 135
0 199 133
0 206 118
0 200 118
1 204 120
0 203 118
0 240 120
1 199 132
1 235 131
1 260 151
1 201 148
1 235 142
0 232 131
0 195 129
1 240 130
1 240 130
1 255 156
1 220 126
0 206 114
0 201 116
1 213 141
1 194 110
1 188 113
0 206 136
0 257 152
0 204 116
0 210 132
RESULT AND SUMMARY
OUTPUT:

1. Description of data:
2. Discriminant function analyses:
The number of discriminant dimensions is the number of groups minus 1. However,
discriminant dimensions may not be statistically significant.

There is two discriminant dimensions, of which is statistically insignificant. The first


F-ratio tests that canonical correlations are non zero;. Since the tests are insignificant,
it follows that dimensions are insignificant and are not needed to describe the
differences between the two groups of person.

The canonical correlations for the dimensions is 0.27.


The discriminant function is:
discriminant_smoking_1 = 1.067*blood pressure - 0.2205*cholesterol

The standardized discriminant coefficients function in a manner analogous to standardized


regression coefficients in OLS regression. For example, a one standard deviation increase on

the blood pressure will result in a 1.0268 standard deviation increase in the predicted
values on discriminant function

The cholesterol will result in a 0.2205 standard deviation decrease in the predicted
values on discriminant function

The canonical structure, also known as canonical loading or discriminant loadings,


represent correlations between observed variables and the unobserved discriminant
functions (dimensions). The discriminant functions are a kind of latent variable and the
correlations are loadings analogous to factor loadings.

S-ar putea să vă placă și