Building discriminant function for repeated measurements data under compound symmetry (CS) covariance structure and applied in the health field

Abstract

Discriminant analysis is a technique used to distinguish and classification an individual to a group among a number of  groups based on a linear combination of a set of relevant variables know discriminant function. In this research  discriminant analysis used to analysis data from repeated measurements design. We  will  deal  with the problem of  discrimination  and  classification in the case of  two  groups by assuming the Compound Symmetry covariance structure  under  the  assumption  of  normality for  univariate  repeated measures data.

The importance of this research represented to find the best model  to classify  a  group of  patients who  suffer  from diabetes.  For  the purpose of studying the effects of  the number of correlations, variances, and umber of  repeated  measurements  on the performance of classification rules for this  type of  data  based on monthly measurements  of  glycosylated  hemoglobin (HbA1C) in the blood was taken in three stages, which  is  the beginning  of  the experiment, and after three months, and  then  six  months for two groups of patients, the first group consists of  (38)  patients  was  suffered  from  diabetes  type (I)  and  the second group includes (33) patients suffered from diabetes type (II).

And through this research, concluded that when the number of parameters began to increase. Thus, the apparent error rate  begin to increasing, and this is what reduces the efficiency of classification rules for this type of data. And  we  recommend  by  using  the linear discriminant function when you focus on the least number of parameters to build the classification rule. And quadratic discriminant procedure Represented by equal the variance and different correlation parameters  under compound symmetry covariance structures