제주 양식넙치 산지가격 예측모형 간 예측력 비교 및 평가

Evaluation and Comparison of the Forecasting Power among Forecasting Models of Jeju Oliver Flounder Producer Price

The purpose of this study is to compare forecasting power among forecasting models of the Jeju aquaculture olive flounder production price and to present models with superior forecasting power. It also aims to help producers and policy authorities make political decisions such as purchasing and stockpiling projects. Accordingly, the ARMA model composed of the Jeju aquaculture olive flounder production price, and the ARMAX and VAR models including the price of Jeju and Wando were used, and the forecasting power among these models was compared by the MDM test. For the analysis, the producer prices of Jeju and Wando were analyzed as stable time series data without a unit root. Through the Granger causality test, the test showed that the production price of Wando in 1 and 2 months ago influenced Jeju price. Based on the redundant variables, heteroscedasticity and autocorrelation tests, ARMA (1,0), ARMA(0,2), ARMA(3,1), ARMAX(1,0), VAR(1) and VAR(2) models were selected as forecasting models for Jeju aquaculture olive flounder production prices. Through the MDM test, the ARMA(1,0) model with the smallest forecasting error and the five models had the same forecasting power in the MSE standard, and the forecasting power of ARMA(1,0), ARMA(3,1) and ARMA(0,2) model was superior in MAE criteria. Therefore, if producers and policy authorities establish management plans and policies related to price forecasting information, this analysis will help alleviate price volatility in the long run.