OIL PRICE FORECAST: A COMPARISON BETWEEN ARIMA, LSTM AND ARIMA-LSTM MODELS

OIL PRICE FORECAST: A COMPARISON BETWEEN ARIMA, LSTM AND ARIMA-LSTM MODELS

• Author
• Cássio Nunes dos Anjos
• Co-authors
• Cássio da Nóbrega Besarria , Rennan Kertlly de Medeiros , Diego Pitta de Jesus , Hyolitta Adrielle Costa de Araújo
• Abstract

• Oil price forecasting is an important and challenging task due to its influence on the world political-economic scenario. In this paper, three models are proposed, the ARIMA Oil price forecasting is an important and challenging task due to its influence on the world political-economic scenario. In this paper, three models are proposed, the ARIMA model which is a widely used statistical method to forecast time series, while the LSTM model is a recurrent neural network that can capture long-term dependencies in sequential data. The third, on the other hand, is an approach that combines the forecasting models to provide a forecast of the WTI oil price.  The WTI oil series is decomposed into its trend and cycle components using the Christiano-Fitzgerald filter and each is predicted separately. With this combination of models and techniques we provide a more accurate and reliable forecast of the WTI oil price to assist in strategic decision making. The data used for the forecast was obtained from the Energy Information Administration and represents a time series with monthly records of the WTI oil price.

• Keywords
• Forecasting models; Oil; ARIMA-LSTM.
• Subject Area
• Econometrics and Numerical Methods

Back Download