Implementation Of The Adaboost Method On Linear Kernel Svm For Classifying Pip Assistance Recipients At SMP Negeri 2 Kejuruan Muda

Abstract

Abstract: This study examines the application of the AdaBoost algorithm to a Linear Kernel Support Vector Machine (SVM) for determining student eligibility for the Indonesian Smart Program (PIP) at SMP N 2 Kejuruan Muda. The main objective is to improve the accuracy and fairness of the PIP aid distribution using advanced machine learning techniques. The dataset used comprises 500 student records, which include demographic, academic, and economic factors. The dataset was divided into training and testing sets, with the AdaBoost algorithm applied to enhance the SVM model’s performance. The study found that the SVM model optimized with AdaBoost was able to classify 91 students as eligible for PIP aid, achieving an impressive accuracy rate of 97.85%. Only 2 students were classified as ineligible, representing 2.15% of the total sample. When compared to the standard SVM model, which also classified 91 students as eligible, the key advantage of AdaBoost lies in its ability to handle borderline data more effectively. AdaBoost improves the classification of students whose eligibility was less clear by reinforcing the importance of difficult-to-classify instances. The model’s higher precision on edge cases indicates that AdaBoost offers a significant improvement over traditional SVM models in handling complex classification tasks. This research concludes that incorporating AdaBoost into SVM models provides a more robust and accurate method for determining student eligibility for government aid programs such as PIP.

Keywords: AdaBoost, SVM, Indonesian Smart Program, PIP aid, machine learning, student eligibility, classification.