Algoritma Backpropagation Neural Network dalam Memprediksi Harga Komoditi Tanaman Karet

Abstract

Rubber plantation sector is one of the leading commodities in East Kalimantan Province contributing greatly to non-oil and gas exports. Currently, the price of rubber in the world is increasingly competitive. The aim of this research is to predict the rubber prices as a reference for the government and companies in making policies and preparing work plans. Data of 60 months during the period of 2014-2018 taken from Plantation office of East Kalimantan Province has been analyzed using Backpropagation Neural Network (BPNN) algorithm in predicting rubber prices. Based on the testing results, parameters of the BPNN algorithm with ratio of 4: 1, architectural models 5-10-10-10-1, trainlm learning function, learning rate of 0.5, error tolerance of 0.01, and epoch of 1000 have gained good accuracy with a mean square error (MSE) of 0.00015464. The results showed that the BPNN algorithm can be used as an alternative method in forecasting.

Keywords

• commodity
• rubber prices
• prediction
• bpnn
• mse