Pendugaan Produktivitas Komoditi Bawang Merah (Allium Ascalonicum L.) dengan menggunakan Machine Learning

Abstract

Productivity estimation is one of the important aspects that can provide information for effective and efficient decision making, especially in the shallot farming sector in the Deli Serdang area. This study aims to predict the productivity value of shallot plants (Allium Ascalonicum L.) using machine learning, especially in Deli Serdang district. In this study, two machine learning algorithms were used; namely Backpropagation Neural Network (BPNN) and Deep Neural Network (DNN) to predict shallot productivity based on historical data which includes climate data, harvest area and production in the Deli Serdang area for the last 10 years. The results show that for the best BPNN prediction value is MAE of 0.638 from the value of harvest area and MAPE of 6.553 from the value of productivity, while for the DNN prediction value, the best result is MAE of 0.575 from the value of harvest area and MAPE of 2.161 from the value of productivity.