- This project explores various deep learning models for time series forecasting of Bitcoin prices, emphasizing the challenges and uncertainties inherent in predicting financial markets.
- Models such as dense neural networks, convolutional and LSTM networks, multivariate approaches, ensemble models, and N-BEATS architecture were trained and evaluated.
- Despite the efforts, the results demonstrate the inherent difficulty in accurately forecasting Bitcoin prices, showcasing the limitations and risks associated with relying on machine learning models in open systems.
- The project emphasizes a critical perspective on financial forecasting and warns against overreliance on models in unpredictable environments.