TIN: TERAPAN INFORMATIKA NUSANTARA
Vol 1 No 1 (2020): Juni 2020

ALGORITMA BACKPROPAGATION DALAM MEMPREDIKSI JUMLAH ANGKA KEMISKINAN DI PROVINSI SUMATERA UTARA

Purba, Roimal Hafizi (Unknown)
Zarlis, Muhammad (Unknown)
Gunawan, Indra (Unknown)



Article Info

Publish Date
30 Jun 2020

Abstract

Poverty is one of the phenomenal problems that Indonesia faces every year. Therefore, this study was conducted with the aim to predict the number of poverty figures by district/city in the province of North Sumatra. The algorithm used to conduct this research is the backpropagation algorithm. This algorithm is one algorithm that is often used to make data predictions. The data used is the data of the poor population in North Sumatra in 2013-2017, which was sourced from the Central Statistics Agency of North Sumatra. Based on this data will be formed and determined the network architecture model used with the Backpropagation algorithm, including 3-9-1, 3-16-1, 3-18-1, 3-23-1, and 3-40-1. From these 5 models after training and testing, it was found that the best architectural model was 3-23-1. The accuracy rate of this architectural model is 97% with an MSE test value of 0.00359. The results of this study are in the form of predictions of the number of poverty in North Sumatra for the next 5 years. The results of this study are expected to be a reference for the regional government of North Sumatra to see the level of development of poverty in North Sumatra for the coming year.

Copyrights © 2020