Penerapan Neural Network dengan Menggunakan Algoritma Backpropagation pada Prediksi Putusan Perceraian

The high divorce rate has a negative impact on couples who will file for and also an extreme children such as psychological disorders of children. magnitude the divorce, it is necessary to predict decision. In this study, application backpropagation method decisions was carried out. data used from Pekanbaru Religious Court 2020 - 2021 totaling 779. dataset obtained not balanced with 724 accepted classes 55 rejected classes, balancing done by reducing excess classes. parameters in study build 3 architectural models [6-7-1], [6-9-1], [6-12-1], learning (0.01, 0.03, 0.09), max epoch sharing (70:30), (80:20), (90:10). results indicate that best model network architecture [6-9-1] 0.09 300 distribution 80% training 20% test accuracy value Mean Squared Error (MSE) 0.1402. successful predicting decisions.