Recomendações para compra e venda de ativos em day trading utilizando redes neurais artificiais.

Abstract

It is notable the increase in the participation of individual investors in the financial market. Many of these investors struggle to accurately discern where to invest their money for higher returns. In this context, the aim of this work was to develop an application to assist in the buying and selling decisions of assets through a supervised neural network trained on data extracted from asset quotations. The developed application seeks to provide an intuitive way to guide daily investment decisions (day trading) regarding which stocks to hold, enter, or exit from each user's portfolio. The principle is to predict the price of the stock for the next day, with a recommendation to sell if the forecasted price is decreasing and to buy if the price is expected to increase. The application also provides an estimate of the profit generated by each buy or sell operation. To achieve this, experiments were conducted with three different assets that matched the major market movements, and it was possible to observe an increase in the initial investment by the user of approximately 23% to 43% in the final amount. The developed tool can assist both novice and experienced users in the stock market.