AI Techniques for Financial Forecasting in the Stock Market

Resumo

https://zenodo.org/records/10694653

In 2020, there was a significant increase in public interest, particularly among young people, in the financial market. This phenomenon is characterized by a growing search for simple and accessible investment options. During the pandemic period, many people turned to the financial market as a means to make profits and plan for the future. A report from B3, the official stock exchange of Brazil, released in December 2020, revealed that approximately 26% of investors are young people between the ages of 18 and 24, as reported by Journal da USP on June 22, 2021.
In this context, this thesis focuses on exploring the LSTM algorithm for forecasting stock closing prices in the financial market. Through the analysis of historical financial data patterns obtained via the Yahoo Finance API (YFinance - Python), this study aims to analyze the potential of LSTM in the predictive analysis of stock prices.
The results obtained demonstrate the algorithm's variable ability to accurately predict the closing prices of stocks of major companies such as Apple Inc., Adobe Inc., and Amazon.com, Inc. In some cases, the predictions closely aligned with the actual data, while in others, they revealed significant discrepancies. These data are important for understanding the limitations and potentials of the employed algorithm and provide a basis for future refinements.