Forecasting Daily Silver Prices Using Arima

Abstract

This study aims to forecast the daily closing prices of silver by employing two distinct approaches: the traditional Autoregressive Integrated Moving Average (ARIMA) model and the modern Extreme Gradient Boosting (XGBoost) algorithm. The dataset covers the period from January 3, 2023, to May 21, 2025, obtained from the Yahoo Finance platform. The ARIMA model was specified using the Box–Jenkins methodology, while the XGBoost model incorporated multiple technical indicators, including simple and exponential moving averages, RSI, and MACD. Model performance was evaluated using Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE). The results indicate that the ARIMA model achieved higher predictive accuracy than XGBoost for this dataset, though XGBoost demonstrated the ability to capture certain nonlinear patterns.