Forecasting Bitcoin returns through time series analysis, emphasizing sentiment analysis on news (using BERT LLM), social media, and Google search trends, with the final model based on Random Forest, augmented with engineered memory features. - View it on GitHub