Previsione del prezzo delle azioni Microsoft: un confronto tra le tecniche allo stato dell'arte ​

Predicting stock prices is challenging due to market volatility and complexity. This research compares ARIMA, Gradient Boosting, XGBoost, and LSTM models in forecasting Microsoft's next-day closing prices. These models, chosen based on relevant literature, were optimized via cross-validation and evaluated over 25 days. Performance was assessed using Mean Squared Error (MSE) and Mean Absolute Error (MAE). Results identify the most accurate model and suggest improvements such as additional features, ensemble methods, and continuous optimization to enhance predictive accuracy in dynamic market environments