Predicting Staple Food Materials Price Using Multivariables Factors (Regression and Fourier Models with ARIMA)

Staple food material prices can be a trending topic in the market. The fluctuation of the price is influenced by many factors. For instance, the weather, oil price, and etc are the external factors of the staple food price. Indeed, the prediction of staple food fluctuation price is important for the farmers, consumers, even government. In this paper, the Linear Regression and Fourier model with ARIMA (Autoregressive Integrated Moving Average) will be used to predict the staple food price which consider the external influences. Here, the results using those two methods are shown in a good agreement with the observation price at market. However, the highest accuracy in predicting price using Fourier regression with ARIMA is obtained for staple food onion which is 96.57%. Meanwhile, using multiple linear regression with ARIMA, the highest accuracy is obtained for staple food red chili with 99.84%. Overall, in this research, Fourier regression with ARIMA is observed better than multiple linear regression with ARIMA method, since the accuracy of Fourier regression with ARIMA is quite stable without disturbance of fluctuation existing data.