PENERAPAN BACKPROPAGATION NEURAL NETWORK DALAM MERAMALKAN PRODUKSI KOPI DI INDONESIA

Abstract

Coffee is one of the most valuable agricultural commodities in the global market, ranking 4th among the ten largest coffee-producing countries in the world. In addition, coffee has the potential to drive the country's economic growth through exports, which can contribute to an increase in national foreign exchange. During the ten-year period from 2014 to 2023, the growth of coffee production was recorded to be lower, with an average increase of about 1.63% per year. The purpose of this research is to determine the forecast of coffee production in Indonesia from 2025 to 2029 using the Backpropagation Neural Network and the accuracy of the method in forecasting coffee production in Indonesia. Data was taken from the Secretariat of the Directorate General of Estates. The method used in this research is the backpropagation neural network method using 4 models of training and testing data, namely 50:50, 60:40, 70:30, and 80:20. Backpropagation Neural Network is a multilayer artificial neural network method that operates in a supervised manner and can be used for classification and forecasting. The results of this study show that the 80:20 model is the best model because the MAPE obtained is 7.672%, with the coffee production forecast in Indonesia for the years 2025 to 2029 being 698,979; 697,202; 696,081; 695,292; 694,700 (tons).With an accuracy level of 7.672%. This value indicates that this method is very good at forecasting coffee production in Indonesian.

Keywords: Coffee, Backpropagation Neural Network, MAPE, Training-Test Data