This is an enthusiastic project about time series analysis. The goal is to forecast the Brent crude oil price for the next 15 days in the case of the SARIMA model, and for the next 30 days in the LSTM model.
Please load the model named "brent_price_forecast_lstm_model" included in this repository to evaluate and use the LSTM model. Due to the inherent randomness of neural networks, it won't be possible to reproduce the exact same model even when using the same hyperparameters and data.
If you will be running a Jupyter Notebook using an IDE like PyCharm, please leave the code cell related to modules importation as it is. On the other hand, if you will be using Jupyter Notebook in its web version, it's recommended to comment lines 28 and 29 of the required modules cell and uncomment line 25 of the same cell.
The data was obtained from Federal Reserve Bank of St. Louis (https://fred.stlouisfed.org/series/DCOILBRENTEU)
Important: Since this is an enthusiast project and not a professional one, the code may not be fully optimized; furthermore, there is potential for significant improvement in both the SARIMA and LSTM models.