Model Prediksi Algoritma ANN Pada Jumlah Ekspor Barang Perhiasan Dan Berharga Menurut Negara Tujuan

Arifah Hanum (Teknik Informatika, STIKOM Tunas Bangsa, Pematangsiantar, Indonesia)
Tri Welanda (STIKOM Tunas Bangsa, Pematangsiantar, Indonesia)
Anjar Wanto (STIKOM Tunas Bangsa, Pematangsiantar, Indonesia)
Agus Perdana Windarto (STIKOM Tunas Bangsa, Pematangsiantar, Indonesia)

TIN: TERAPAN INFORMATIKA NUSANTARA
Vol 3 No 1 (2022): June 2022

Currently Indonesia is one of the exporting countries to industrialized and developing countries. The methods carried out in the research of the prediction of the export of jewelry and valuables from this main destination country use the ANN (Artificial Neural Network) method. The research data used comes from the official website of the government, the Indonesian Central Statistics Agency. In this study, the data used is data from 2013 to 2020 consisting of 8 destination countries, namely Switzerland, Singapore, Hong Kong, United Arab Emirates, South Africa, Taiwan, the United States, and India. Based on this data can be determined network architecture model, namely 3 - 4 - 1, 3 - 8 - 1, 3 - 12 - 1, 3 - 16 - 1 and 3 - 20 - 1. After training and testing of the 5 models, it can be obtained that the best architectural model is on the 3-12-1 model with an MSE value of 0.033777975 on the ANN method