A smart and intelligent generation is needed to develop a Nation and State. One of important factors to create this generation is education. However, there are still many students with great skill and potential who can not continue their study at school because they are financially disadvantaged. Fortunately, there are also many of them who earn scholarship. Yapimda has donation both from the government in the form of BOS (School Operational Aid) and from other parties. Yet, there is no exact measurement to determine the eligibility of students to be awarded scholarship. In education environment, especially at school, there must be some regulations or classifications in determining students to be given scholarship. Therefore, this research conducts an algorithm comparison between Multilayer Perceptron (MLP) and Support Vector Machine (SVM) to be applied to the data of students receiving scholarship. The purpose of the research is to measure the accuracy level of 2 algorithms compared in selecting potential recipients of scholarship in SMK YAPIMDA Jakarta. The test result by measuring the performance of the two algorithms using a Cross Validation test method, Confusion Matrix and ROC Curve, shows that algorithm of Multilayer Perceptron (MLP) has the highest accuracy of 85.82%, while algorithm of Support Vector Machine (SVM) has the lowest accuracy of 83.98 %.
Copyrights © 2017