Predicting effect on student s enrollment in higher education using data mining techniques

Abstract

The enrollment trends in higher education institutions play a pivotal role in newlineshaping educational policies and resource allocation. This research focuses on newlinepredicting the effects on student enrollment in universities in Haryana, India, newlineutilizing data mining techniques. We have collected extensive data from various newlineuniversities in Haryana, encompassing a wide array of variables that influence newlineenrollment. newlineOur research employs a two-step approach to analyze and predict student newlineenrollment. Firstly, we conduct descriptive analysis using IBM SPSS Statistics newlineto gain insights into the data, identify trends, and understand the key factors newlineaffecting enrollment. Subsequently, we employ machine learning algorithms newlineand prediction models, utilizing the Weka toolset. To predict student enrollment newlineaccurately, we evaluate the performance of multiple classifiers, including J48, newlineJRIP, IBK, and Naive Bayes. These classifiers enable us to capture the complex newlinerelationships within the data and make precise enrollment forecasts. newlineFurthermore, we construct an ensemble model that combines the strengths of newlinethese classifiers, achieving an impressive prediction accuracy of 93.22%. This newlineresearch contributes to the field of education by providing a robust predictive newlinemodel that can assist educational institutions and policymakers in making newlineinformed decisions regarding resource allocation, capacity planning, and other newlinecritical aspects related to student enrollment management. Additionally, it newlinehighlights the potential of data mining techniques in improving the newlineunderstanding and management of enrollment dynamics in higher education newlineinstitutions. newline