Comparison of the efficiency of principal component analysis and multiple linear regression to determine students' academic achievement

Abstract

The Georgia Ministry of Education and Science is responsible foundation to prepare the National Unified Entrance Examination (NUEE) in Georgia. Georgian Language, Logic, English Language and Mathematics are some of the categories of this examination. In this study we focused on how NUEE affects the grade point averages (GPA) of the students of International Black Sea University (IBSU). The relation between NUEE scores and GPA is represented and compared for the all students of the faculty of Computer Technologies and Engineering (CT&E) and the faculty of Business and Management (B&M). The research is also done and indicated separately for female and male students. The major purpose of this study is to compare the efficiency of multiple linear regressions (MLR) and principal component analysis (PCA) in predicting the response variable GPA using NUEE's explanatory variables (X). In the consequence, using principal components as entries improves multiple linear regression prediction by reducing complexity and high dimensionality.