Is Facebook PROPHET superior than hybrid ARIMA model to forecast crude oil price?
Oil price forecasting has received a great deal of attention from practitioners and researchers alike, but it remains a difficult topic because of its dependency on a variety of factors, including the economic cycle, international relations, and geopolitics. Forecasting the price of oil is a dif...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Penerbit Universiti Kebangsaan Malaysia
2022
|
| Online Access: | http://journalarticle.ukm.my/20468/ http://journalarticle.ukm.my/20468/1/22.pdf |
| Summary: | Oil price forecasting has received a great deal of attention from practitioners and researchers alike, but it remains a
difficult topic because of its dependency on a variety of factors, including the economic cycle, international relations,
and geopolitics. Forecasting the price of oil is a difficult but gratifying task. Motivated by this issue, we present a robust
model for accurate crude oil price forecasting using ARIMA and Prophet models based on machine learning technique
to produce a reliable weekly and monthly crude oil price predictions. We apply the Savitzky–Golay smoothing filter to
get a better denoising performance for our forecast models. For model evaluation, we apply cross validation with sliding
windows on both models and compares the performances using RMSE and MAPE. The results show that the ARIMA-based machine learning approach performs better as compared to the Prophet model for both one-week and one-month
forecast ahead intervals. |
|---|