Deep Learning-Enhanced Intraday Stock Trade Price Prediction

Investors have shown a lot of interest lately in cryptocurrencies due to their decentralization and irreversibility. Forecasting the worth of digital assets like Bitcoin is challenging because of their instability; therefore, establishing an efficient investing plan requires accurate price predictions. With this in mind, the proposed system implemented a modern approach for forecasting Bitcoin prices, using change point detection to segregate time-series data and enable isolated normalization. Predicting stock market shares also involves foreseeing a company’s future financial stocks, for which machine learning and its models are an emerging technology. Besides, this work included the Satoshi Sling Algorithm (SSA) Ensemble Models which rely on self-attention components to forecast stock values; date, entry, low_value, high_value, exit, adj_exit, size, and sentiment being relevant elements in such predictions. Tested with real BTC figures and different method set-ups showed promising outcomes regarding the Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Mean Squared Error (MSE), Mean Poisson Deviance (MPD), and Mean Gamma Deviance (MGD), surpassing K-Nearest Neighbor (KNN), and Random Forest Regressor (RFR).