ADS Capstone Chronicles Revised
23
10.0.3 Enhance Feature Engineering Continuously exploring and integrating new technical indicators and features can improve model performance. Features such as order book data, trading volumes, and on-chain metrics could provide deeper insights, keeping the field of cryptocurrency price prediction dynamic and engaging, which is essential for the audience to feel connected to the field. 10.0.4 Specialized Models for Bitcoin Given Bitcoin ’ s unique market behavior, developing specialized models or incorporating additional features specifically for BTC-USD could address its forecasting challenges. and regularization techniques should be applied to mitigate potential overfitting. Overfitting can lead to models that perform well on historical data but poorly on new data, so methods such as cross-validation and grid search should be employed to refine model parameters continuously and ensure the model ’ s generalizability. 10.0.6 Explore Advanced Models While Ridge Regression performed well, exploring other advanced models, such as neural networks and hybrid models, could improve performance. Models like LSTM (Long Short Term Memory) and Transformer-based architectures are up-and-coming for time series forecasting. 10.0.5 Regularization and Hyperparameter Tuning Further hyperparameter tuning 10.0.7 Automate Data Extraction and Forecasting As a future enhancement, it is recommended that data extraction from Coinbase be automated and integrated into a database. Once housed within
the database, the data would undergo rigorous cleaning and preprocessing. Subsequently, with its robust algorithms, the model would be programmed to execute automatically, forecasting future values on a predefined schedule. Employing Airflow to orchestrate this automation would streamline operations, significantly improving efficiency and ensuring consistent, timely updates. By implementing these recommendations, future studies can build on this project ’ s findings and achieve more accurate and reliable cryptocurrency price predictions. Integrating diverse data sources and advanced modeling techniques will be crucial in navigating the complexities of the cryptocurrency market. Aghabozorgi, S., Shirkhorshidi, A. S., & Wah, T. Y. (2015). Time-series clustering - A decade review. Information Systems, 53, 16 – 38. https://doi.org/10.1016/j.is.2015.04.007 Bhatt, S., Ghazanfar, M., & Amirhosseini, M. (2023). Machine learning-based cryptocurrency price prediction using historical data and social medias. 5th International Conference on Bollen, J., Mao, H., & Zeng, X. (2011). Twitter mood predicts the stock market. Journal of Computational Science, 2 (1), 1 – 8. https://doi.org/10.1016/j.jocs.2010.12.007 Brownlee, J. (2017). Introduction to time series forecasting with Python: How to prepare data and develop models to predict the future. Machine Learning Mastery. References Machine Learning & Applications (CMLA 2023).
90
Made with FlippingBook - Online Brochure Maker