.

The research is aimed at applying discriminant analysis in developing a model for the prediction of the final year graduating class of degree and also to determine the influence of hundred and two hundred levels student’s results on the final graduating class of degree in the Department of Mathematical Sciences, Federal University of Technology, Akure, Ondo State, Nigeria. Data were collected by taking a random sample of 150 academic records of alumni students and categorized into three groups as historical data in the department of Mathematical Sciences. Fisher’s discriminant analysis were used to analyze the collected data that resulted in two eigenvalues
(
) and corresponding eigenvectors. The first discriminant function associated with
coefficients (
) revealed positive contribution for the variables: first semester result, hundred level (
=
), two hundred level (
= 0.612), second semester result, hundred level (
= 0.417) and two hundred level (
= 0.578). The second function which is not significant revealed a negative contribution for the variables. This revealed that these variables highly contribute to the final class of graduation. The models obtained were able to classify different groups in the historical data up to 97.33% (hit ratio). The function was a valid tool for classifying students of unknown group.
Key words: Multivariate normality test, Discriminant Analysis, Wilks’ Lambda test, Classification matrix, Press Q Statistic.