Author(s): Teddy Oswari, Tristyanti Yusnitasari, Reni Diah Kusumawati, Irvan Setiawan
Agriculture is a sector that has great influence and potential to be exploited for the Indonesian economy. The Indonesian agricultural sector is considered to be very influential on the national economy based on the main macroeconomic variables in the form of the composition of the workforce and the price index received by farmers (IT). Most of the workforce in Indonesia, especially for the small group, works in the agricultural sector. The price index received by farmers (IT) is a value that shows the level of development of farmer production. During the COVID-19 pandemic, Indonesia's agricultural sector is considered to still have a role in national economic growth. Even so, there are still impacts from the COVID-19 pandemic on Indonesia's agricultural sector. One of them is the disruption of farmers' production in all regions. Therefore, one solution to maintaining the stability of national economic growth in the agricultural sector is to predict the development of food crop farmers' production. In this research, prediction or forecasting will be carried out with the Auto Regressive Integrated Moving Average (ARIMA) algorithm with parameters SARIMA(2, 1, 2) x (0, 1, 1, 1) and the Long Short Term Memory (LSTM) algorithm with LSTM parameters 100, dropout 0.2, 100 times. The results of the forecasting analysis of the two models show that the LSTM model has more accurate prediction results than the ARIMA model because the MSE value in the LSTM model is lower (0.1051) than the ARIMA model (0.2692).