Penerapan Algoritma Jaringan Saraf Tiruan Metode Backpropagation untuk Memprediksi Jumlah Nilai Ekspor di Provinsi NTB

Abstract

This paper presents the application of the Backpropagation method of the Artificial Neural Network algorithm in the case study to forecasting the amount of export value in NTB province. This forecasting process uses two scenarios, namely forecasting the amount of export value in NTB province and forecasting the amount of export value based on a commodity which then the forecasting results based on these two scenarios will be compared. Based on the results of the system testing that has been done, the best network architecture is obtained from 12-4-1, the best value of learning rate is 0.2 and the best number of epochs is 6000, which in the training device produces these variables resulting in an MSE value is 0,0034 and MAPE value is 8.52% and for the testing result MSE value is 0.0169 and MAPE value is 17.94%. Based on the results of forecasting with two scenarios that have been carried out are forecasting results that are negative. This is because the pattern of data used is not stable so that it can produce negative values.