Budiharjo, Triyuni Soemartono, Agus Perdana Windarto, Tutut Herawan
This study aims to predict factors causing tuition fees payment problems with artificial neural networks using Backpropagation algorithm. The data is collected by interview and questionnaire from students having problem in tuition fee payment of a private higher learning institution in Indonesia where in total 20 sample data are obtained. There are six categories and in predicting these factors, the Backpropagation algorithm of artificial neural network has been employed involving training and testing. The training stage is used to get the best architectural model from a series of experiments. Meanwhile, the testing stage is used to see the accuracy of the model used. Four architectural models i.e. 6-5-1, 6-10-1, 6-5-10-1 and 6-10-5-1 have been used. The results show that, the best architectural model is achieved for 6-5-10-1 model with Mean Square Error 0.0010002075 and its respective accuracy up to 80%. © 2018 SERSC Australia.
Universitas Prof Dr Moestopo (Beragama), Jakarta, Indonesia; STIKOM Tunas Bangsa, Pematangsiantar, Indonesia; Universitas Negeri Yogyakarta, Indonesia; Universitas Teknologi Yogyakarta, Indonesia; Politeknik Negeri Malang, Indonesia
