Prediction Techniques Applied to Gold Price Analysis: A Comparative Approach

In this study, we address the challenge of developing an accurate and reliable model to predict the price of gold. Our main objective is to analyze and compare different approaches, such as Linear Regression models, Support Vector Regression, and Regression Trees, to determine which one offers the best gold price prediction accuracy. Through an exhaustive bibliographical review and the analysis of historical data, we seek to identify patterns and trends that allow us to improve the predictive capacity of the model. In the development process, we have used a rigorous and transparent workflow, from the definition of the problem to the evaluation of the models by dividing the data into training sets, cross-validation, and testing. Our methodology includes collecting data from trusted sources and preparing the data through proper cleaning, exploration, and transformation. The results obtained so far show that the linear regression model reaches an R2 metric greater than 0.89 but with an RMSE of 31.38 and an MAE of 17.84. Although this approach is simple and easy to understand, we need to explore other machine-learning techniques to improve the accuracy of gold price prediction. The findings of this study will have significant implications for both investors and the industry, providing valuable information for informed gold investment decision-making and risk management, as well as for production planning and optimization of resources in the mining industry.