1. Discriminant Analysis
DA is a technique for analyzing data when
the criterion or dependent variable is
categorical and the predictor or
independent variables are interval in
nature.
2. Discriminant Analysis
DA is sometimes also called
Discriminant factor analysis
Canonical discriminant analysis
3. Objectives
Development of discriminant functions
Examination of whether significant
differences exist among the groups, in
terms of the predictor variables.
Determination of which predictor
variables contribute to most of the
intergroup differences
Evaluation of the accuracy of
classification
4. Multiple Discriminant Analyses
Discriminant functions
A function like multiple regression but
different in the aspect that the dependent
variable is categorical.
5. Discriminant Analyses
The discriminant score , or DA score, is
the value resulting from applying a
discriminant function formula to the data
for a given case.
6. Multiple Discriminant Analyses
Discriminant coefficients are the regression-
like b coefficients in the discriminant function, in
the form D = b1x1 + b2x2 + ... + bnxn + c,
where D is the latent variable formed by the
discriminant function, the b's are discriminant
coefficients, the x's are predictor variables, and c
is a constant.
7. Multiple Discriminant Analyses
The structural discriminant function coefficients
are partial coefficients, reflecting the unique
contribution of each variable to the classification
of the criterion variable
The standardized discriminant coefficients , like
beta weights in regression, are used to assess
the relative classifying importance of the
independent variables
1 st table. A sample of 34 values wre taken and all of em were analyse 2 nd table Gives basic descriptive of each behaviour Valid n -- all 20 implies all 20 cases have been used.ie it gves number of variables used category wise.and no spoecific wwts have been assigned to any value 3 rd table Log determinant is for grup 0 and 1 for all the variables matyrix is drawn…and log determinant shud be higher 4 th table Sig value of F shud be less than 0.05 its d confidenc levels Eigen values shud be greater than 1 …for goodness of fit If EV >1for good discriminating power Wills lambdabetween 0 and 1…closer to 0 better it is Standardized function…relative importance of variables is loooked into..ie higher the value higher the importance Structure matrix and canonical are used in making overall function ie -6.058+0.032age+0.222income……………………… Grup centroid gives the avg of all d values of each grup Cut off score is weighted avg of grup centoid ie {20(1.914)-14(2.734)}/34 = 0.2 this is the ecentroid Nw if theD value is <0.2 then grup 0 ifD value>0.2 then grup 1 Clasification stats gives hw functions have been classified Squared mahalonobis distanceto centroid gives the values of all values of the distances of all variables from their own centroid..this distance is reported in further analysis df – degree of freedm