Advantages of Discrimininant Analysis
- Multiple dependent variables
- Reduced error rates
- Easier interpretation of Between-group Differences: each discriminant function measures something unique and different.
Disadvantages of Discriminant Analysis
- Interpretation of the discriminant functions: mystical like identifying factors in a factor analysis
- Each discriminant function formed is distributed normally in each group being compared.
- Each discriminant function is assumed to show approximately equal variances in each group.
- Patterns of correlations between variables are assumed to be equivalent from one group to the next
- The relationships between variables are assumed to be linear in all groups
- No dependent variable may be perfectly correlated to a linear combination of other variables (Multicolinearity)
- Discriminant analysis is extremely sensitive to outliers.
Pitfalls of Application of Discriminant Analysis in the Business, Finance and Economics
Of the most applications of the Discriminant Analysis that were applied to the business, finance and economic literature most of them have suffered from methodological and statistical problems that have restricted their practical usefulness of their results. Though statistical problems are not unique to economics and finance but the nature of the subject matter and the data are such that one expects the more frequency of encountering statistical difficulties than other area of application.
The several difficulty types are as follows:
1. The distribution of variables
2. The group dispersions
3. The interpretation of significance of individual variables
4. Reduction of dimensionality
5. The definition of the groups
6. The choice of appropriate apriori probabilities and/ or cost of misclassification
7. The estimate of classification error rates
Vol. 32, No. 3, Jun., 1977
Author: Prashansa Wankhede
Finance Group 5