Machine Learning Models for IBOVESPA Stock Price Trend Prediction

Abstract

Financial markets perform a fundamental role in the economic organization of countries, motivating investors to seek improvements in technical analysis tools to optimize gains. This study applies several Machine Learning techniques to predict the trend (up, down and lateral) of the IBOVESPA index on a weekly interval between technical indicators (explanatory variables). Statistical parameters such as precision, accuracy, ROC curve and AUC were evaluated, highlighting the performance of the KNN, Random Forest and Logistic Regression models. It is concluded that Machine Learning techniques are effective in the investment sector, offering impressive results for the market and future research.