Wavelet 분해를 이용한 단기 천연가스 가격 예측

초록

Although natural gas contracts have been done based on oil prices, many recent research argued that the natural gas and oil prices are decoupled. Therefore, it is necessary to analyze the natural gas price as an independent price. In the light of this, we forecasted the short term natural gas price using wavelet decomposition method. In order to compare a forecasting power, three approaches are applied. The first approach is ARIMA only case. We applied an ARIMA into the original data. The second one is wavelet-ARIMA case. In this case, ARIMA is applied into the approximation. In the third approach, we implement wavelet-GARCH analysis. In this approach, an approximation is treated same as second case, but GARCH is applied to details. As results, the wavelet-applied approaches are better than only ARIMA case for the forecasting power. Consequently, we could confirm the usefulness of wavelet-applied approaches in natural gas price forecasting.

키워드