The Role of International Variables in Predicting Gold Prices: Analysis with Machine Learning Algorithms

Abstract

This study predicted gold prices using the S&P 500, crude oil prices, dollar index and volatility index variables and various machine learning methods. Research results show that gold prices are predicted successfully with existing methods. According to analysis, the most successful gold price forecasters are the WTI, VIX, S&P 500 and US dollar indexes. The machine learning method that best predicts gold prices is the random forest method, with an R-square of 0.96 and a MAPE value of 3.5%. This study is expected to contribute to the literature in measuring the success of machine learning algorithms in price prediction and the predictability of gold prices within the framework of the efficient markets hypothesis.