Search

Forecasting food prices has been developed by several researchers by utilizing various algorithms to forecast over the next few months. However, most studies focus on univariate forecasting methods and underutilized factors that might be predictive of price dynamics. The recent COVID-19 pandemic presented another factor that can significantly affect the dynamics of food commodities prices. This study aims to carry out multivariate forecasting of basic food commodity prices considering the economic and health factors such as inflation rates, fuel prices, the exchange rate of rupiah against the dollar, and the number of positive cases of Covid-19 in Jakarta with the Long-Short Term Memory algorithm were to test accuracy forecasting results based on MAPE values. This research consists of five steps: literature study, data pre-processing, the architectural parameters initialization of the LSTM model, the LSTM model analysis, and the LSTM model evaluation. The results of the LSTM model for each basic food commodity produced an average MAPE value of 0.91501%. Predictor variables with the greatest influence on food commodity prices in order are the pattern of movement in food prices themselves, Fuel RON 92, exchange rates Rupiah against the US Dollar, inflation rate, and the number of Covid-19 active cases. [ABSTRACT FROM AUTHOR]