Investigating Admission Process of Grade 12 Students by Higher Education Admission Center (HEAC) to Sultan Qaboos University

Abstract

In this paper, we mainly focused on one of multivariate statistical methods called Discriminant Analysis DA. It has two main different approaches: Linear Discriminant Analysis (LDA) and Quadratic Discriminant Analysis (QDA) including k-Nearest neighbors (K-NN). The main objective of our study was to investigate and predict the performance (success or non-success) of students grade 12 in college of science at Sultan Qaboos University using DA. Totally of 1360 students were taken as sample size and seven predictor have been used. We assessed the predictive power of these methods by computing the Apparent Correct Classification Rate (ACCR). By exploring our dataset with a help of chi-sqaure test, we got that there was significant difference in the performance between male and female. Moreover, we found that Quadratic Discriminant Analysis (QDA) is the most appropriate model for predicting the performance of female students with (ACCR) equal to 95%. On the other hand, (k-NN) was found to be the most suitable statistical method for predicting the performance of male students with (ACCR) equal to 62%.

Keywords: Discriminant Analysis, Apparent Correct Classification Rate, Predictive power