Authors : Agus Widodo, Heni Kusdarwati and Samingun Handoyo
Abstract: The high consumption of garlic in Indonesia cannot be followed by an adequate supply of garlic, so that, until now Indonesia still imports garlic. Policy formulation in the field of essential goods requires accurate information The study discusses modeling stages and develops of the average prices of garlic in Indonesia with the ARIMA Box-Jenkins method. The monthly data of average prices of garlic in the January 2014 to June 2018 period obtained from the directorate of basic goods and essential goods of the Ministry of Trade of the Republic of Indonesia. Model development uses data from the period of January 2014 to December 2017 which were referred to as in-sample data while data from January 2018 to June 2018 period as out-sample data to evaluate model performance. Modeling begins with the transformation of time series data into stationary time series data, identification pattern of common models via. plot of ACF and PACF values, estimating and testing of parameters to obtain a feasible model, then continued with a model diagnostic test and finally, gets the best model is ARIMA Model (1, 1.0). Based on the best model, the comparison plot was made between the actual and predicted values that were resulted by both of in-sample and out-sample data and also was calculated the coefficient of determination (R2) as accuracy measure. The actual and predicted values plot in the in-sample data are very coincident and produce R2 which is quite high at 91. 4%. However, the performance of the ARIMA (1, 1, 0) Model is moderate when forecasting out-sample data whicch R2 is 59.8%.
Agus Widodo, Heni Kusdarwati and Samingun Handoyo, 2019. Modeling Average Prices of Garlic in Indonesia. Journal of Engineering and Applied Sciences, 14: 7943-7950.