A REVIEW ON DATA MINING AND MACHINE LEARNING METHODS FOR STUDENT SCHOLARSHIP PREDICTION

Abstract

Student scholarship prediction has become an important application in educational data mining, aimed at identifying deserving candidates based on academic performance, socio-economic factors, and behavioral attributes. With the increasing volume of student data generated by educational institutions, traditional manual evaluation methods are becoming inefficient and prone to bias. This review paper focuses on analyzing various data mining and machine learning techniques used for predicting student eligibility for scholarships. The study examines methods such as Decision Trees, Random Forest, Support Vector Machines (SVM), Naïve Bayes, and Neural Networks, which have been widely applied to classify and predict student outcomes. The review highlights the role of data preprocessing techniques, including data cleaning, feature selection, and handling missing values, in improving model performance. It also discusses the importance of selecting relevant features such as academic scores, attendance, family income, and extracurricular activities. Comparative analysis of different algorithms shows that ensemble methods like Random Forest often provide higher accuracy, while simpler models like Naïve Bayes offer faster computation with reasonable performance. Additionally, the paper explores challenges such as data imbalance, privacy concerns, and the need for interpretability in decision-making. Overall, this review provides insights into the effectiveness of machine learning approaches for scholarship prediction and emphasizes the need for robust, fair, and transparent models. Future research directions include the use of deep learning, hybrid models, and real-time data analytics to enhance prediction accuracy and support educational decision-making processes.