Artificial Neural Network Prediction Model for Agricultural Commodity Production Using Backpropagation Algorithm

The development of Artificial Intelligence (AI) technology has been widely used by the Government and Society to support daily activities, including supporting the decision-making process. In Indonesia's agricultural sector, innovations are currently being implemented using Machine Learning methods, especially Artificial Neural Networks, to estimate the yield of an agricultural commodity. This technology is very relevant to be applied in the agricultural sector, especially since the majority of Indonesians are farmers. With prediction of production and prices, the Government can estimate the amount of production and immediately set a strategy to keep prices stable. The use of predictive data on agricultural production results is very important in maintaining food availability and preventing price fluctuations that affect society. This study uses data on chili commodities, employing a qualitative method with the Backpropagation Algorithm of Artificial Neural Networks. The objective is to generate projections of the Artificial Neural Network (ANN) model using the Altair AI Studio with minimal error so that better prediction values and performances are produced. Based on the results obtained, the best network architecture is the 12-25-1 model for large chili production, and 12-15-1 for bird’s eye chili pepper. This model is proven to be able to help production planning, supply distribution arrangements, and maintain price and supply stability by related agencies.