Indonesian Food Price Prediction with Adaptive Neuro Fuzzy Inference System

Abstract

Rising food price in Indonesia is an uncertain issue that significantly affects the lives of the people and the country’s economy. In addition, the demand for Indonesian staple foods, namely rice, is very high. Of course, the fluctuations in the price of rice will have a significant impact on the Indonesian population. So, we need a system that functions to predict food prices. Therefore, this study aims to predict Indonesian food prices. Rice price data is taken from the Central Bureau of Statistics (BPS) website and the prediction was built using the Adaptive Neuro-Fuzzy Inference System (ANFIS) method. Then the model will be evaluated using Mean Squared Error (MSE) and Mean Absolute Percentage Error (MAPE). In this study, the ANFIS model is made with the number of membership functions (MFs) as much as 2, the number of iterations (epochs) is 300, and the type of membership function (MF type) is gaussmf. The training and testing results using ANFIS show that the MSE and MAPE values are 0.9176 and 0.70059%. These results prove that the ANFIS method can predict Indonesian food prices well.