Gold Price Relative Return Prediction with Machine Learning Models

Runjie Zhang

2024

Abstract

Gold is a safe haven asset during the crisis; it can help investors to hedge against inflation and economic uncertainty. Thus, predicting gold return is essential for financial institutions and individual investors. This paper uses Extreme Gradient Boosting (XGBoost), Support Vector Regression (SVR) and Random Forest (RF) model to predict gold return—dataset sources from Yahoo Finance. Features of oil price, volatility index, S&P 500 index, and USD index are add-ed for better prediction. Technical features such as MACD difference, RSI, and Bollinger%B are applied for better accuracy. To find the best parameters, grid search is conducted. To eval-uate the model's performance, mean square error, root mean square error, mean absolute error, R-squared (R2) value, and trend accuracy are calculated and compared among models. RF and SVR give an R2 value of 0.79, and XGBoost gives an R2 value of 0.72. The overall perfor-mance of the SVR and RF models is nearly the same, but the RF model has higher trend accu-racy and better prediction fitness. The SVR model performs much better in predicting extreme values than RF.

Paper Citation

in Harvard Style

Zhang R. (2024). Gold Price Relative Return Prediction with Machine Learning Models. In Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML; ISBN 978-989-758-754-2, SciTePress, pages 579-584. DOI: 10.5220/0013528700004619

in Bibtex Style

@conference{daml24,
author={Runjie Zhang},
title={Gold Price Relative Return Prediction with Machine Learning Models},
booktitle={Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML},
year={2024},
pages={579-584},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013528700004619},
isbn={978-989-758-754-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML
TI - Gold Price Relative Return Prediction with Machine Learning Models
SN - 978-989-758-754-2
AU - Zhang R.
PY - 2024
SP - 579
EP - 584
DO - 10.5220/0013528700004619
PB - SciTePress