Enhancing College Admission Decision-Making Using Machine Learning

Abstract

This research explores the use of machine learning techniques to enhance the college admission process. The use of traditional frameworks in evaluating student applications for college admissions has been heavily criticised for their inefficiencies, subjectivity, and disparities. This research addresses these challenges by proposing using machine learning to aid the admissions departments of colleges in Ghana in making their decisions. The study investigates two research questions: (1) How do different machine learning models perform in predicting college admissions for Ghanaian students? (2) Which part of the college admissions process can be effectively automated using machine learning? The results show that the three algorithms (multiclass regression, multilayer perceptron, and random forest classifier) implemented performed well in predicting admissions outcomes; however, they should not be used as a standalone, but they could be used to work together with the admissions department to aid in college admission predictions. The study also showed that machine learning could automate the processing of students' grades with minimal human effort, significantly reducing the workload of admissions officers and ensuring consistency in the evaluation process. This was done by building an Automated Student Grade Processing and Scoring System as proof of concept.