Classifying beneficiaries of islamic boarding school rehabilitation aid based on neural network approaches: A case of the religious affair ministry of East Java, Indonesia

Abstract

Islamic Boarding Schools (pesantren in Indonesian language) often need government funding grants for improving education services, i.e. rehabilitation aid. Many affecting variables such as the number of student, pesantren activity type, and infrastructure condition need further examination, in addition to the large number of institutions. Because of those complex variables and the absence of definite variables pattern about correlation with the target classes, this research proposed two neural network based model for classifying beneficiaries to determine rehabilitation aid for the pesantren institutions and compared which is the best. 15 input variables were used as the features in learning model are accordance with 4 target classes. Neural Network formed from the learning process can generate new data classification as much as 100% for Backpropagation with accuration value 0.5, and 94.45489% for Radial Basis Function with accuration value 0.428571429.