Are Silver Prices Predictable? A Machine-Learning Study with High Predictive Accuracy

Sanjay Bharath

Abstract: This study examines the predictability of silver prices using time series and machine learning models. Daily silver price data from 2000 to 2025, comprising 6,360 observations, were analysed using ARIMA, Random Forest, and XGBoost models. The dataset was split chronologically into 70% training and 30% testing sets. Model performance was evaluated using RMSE, MAE, and MAPE. Results show that the ARIMA model achieved the lowest prediction errors, indicating strong performance in capturing short term price dynamics. Machine learning models were able to follow general trends but showed reduced accuracy during periods of high volatility. These findings suggest that traditional time series models remain effective for short term forecasting of silver prices, although incorporating additional explanatory variables may improve machine learning performance. The study highlights the challenges of forecasting volatile commodity markets and suggests directions for future research.

Keywords: Predictive accuracy, ARIMA, XGBoost, Random Forest, Decision Trees, Time series forecasting, Volatility modelling

How to Cite?: Sanjay Bharath, "Are Silver Prices Predictable? A Machine-Learning Study with High Predictive Accuracy", Volume 15 Issue 4, April 2026, International Journal of Science and Research (IJSR), Pages: 832-841, https://www.ijsr.net/getabstract.php?paperid=SR26404180625, DOI: https://dx.dx.doi.org/10.21275/SR26404180625

Download Citation: APA | MLA | BibTeX | EndNote | RefMan