Applications of discriminant analysis in medical diagnosis

In this paper, the discriminant analysis is used to classify the most wide spread heart diseases known as coronary heart diseases into two groups (patient, not patient) based on the changes of discrimination features of ten predictor variables that we believe they cause the disease. A random sample for each group is employed and the stepwise procedures are performed in order to delete those variables that are not important for separating the groups. Tests of significance of discriminant analysis and estimating the misclassification rates are performed.