Projeto de robô trader assistido com inteligência artificial
Carregando...
Arquivos
Data
2021-11-24
Autores
Sassaki, Bruno Hiroyuki
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Desde 2018, a taxa básica de juros da economia, ou taxa Selic, foi gradativamente reduzida. Com essa constante diminuição uma parcela do povo brasileiro passou a se interessar cada vez mais em buscar alternativas para a manutenção de um rendimento aceitável, bem como a melhoria de sua condição atuando no mercado financeiro. Entretanto, apesar de ser uma opção muito atrativa para essa população, não valia a pena correr os riscos que esse mercado oferece em troca de bons rendimentos no cenário anterior de altas taxas de juros. Simultaneamente, no período mencionado, o número de CPFs cadastrados na Bolsa de Valores Brasileira aumentou de forma exponencial. Isso se deve não apenas à busca de alternativas de rendimentos, mas também aos influenciadores digitais que ajudaram a expor os conceitos financeiros de maneira simplificada a fim de amadurecer a visão desse segmento de pessoas que tenta investir parte de sua renda. Uma forma de melhorar a rentabilidade desses investimentos é possível a partir da utilização de robôs traders, que por meio de análises de gráficos tentam identificar padrões, e assim obter lucro nas operações baseadas em tomada de decisão, pela identificação desses padrões. Inicialmente, para verificar o potencial de um robô assistido por inteligência artificial, foi desenvolvida uma estratégia financeira levando-se em conta o gerenciamento de riscos das operações que serão executadas pelo robô. A estratégia foi implementada com o software Metatrader 5, utilizando a linguagem de programação MQL5, que permite a realização de operações na bolsa de valores em tempo real, bem como os testes de possíveis operações no passado, para simular o comportamento que poderá ser alcançado no mercado. Com o código definido passou-se à aplicação de otimização da estratégia de gestão de riscos, por meio da implementação dessa inteligência artificial, proporcionando operações com maior precisão. Os resultados obtidos, indicam que o menor percentual de otimização foi de aproximadamente 1%, resultando em um saldo de carteira de R$ 68.200,00, comprovando, assim, a eficiência da adição do algoritmo genético ao algoritmo implementado. Destaca-se ainda que nas ocasiões em que foram observadas situações anteriores de prejuízo, passou-se a obter lucro. Cabe ressaltar que a performance foi melhorada não apenas pelo resultado de carteira, mas também com menor exposição ao risco, uma vez que se efetivou menor quantidade de operações, e com maior precisão.
The financial market has always been very attractive for the Brazilian population, however with the high interest rate it wasn’t worth the risk it offers in exchange for good financial returns. With the constant decrease in the base interest rate (Selic), Brazilian people started to seek alternatives to maintain an acceptable income and thus maintain or improve their financial condition. The number of CPFs registered at Brazilian’s stock market has been increasing exponentially in recent times, and this is due not only to the search for income alternatives, but also to digital influencers, who try to bring financial concepts in a simplified way in order to mature the vision of Brazilian population. One way to improve profitability is with the use of robot traders, which through chart analysis try to identify patterns, and thus obtain profit in operations based on decision making, identified by these patterns. To identify the potential of a trading robot assisted by artificial intelligence, initially a financial strategy was developed taking in account a risk management of the operations that will be executed by the robot. The defined strategy was implemented in the Metatrader 5 software, using the MQL5 programming language, which allows the performance of stock exchange operations in real time, as well as the tests of possible operations in the past, to simulate the behavior that can be achieved in the market. Once the code has been defined, it is now possible to apply optimization through the implementation of artificial intelligence, which should improve the risk management strategy, providing more accurate operations. The results obtained prove the efficiency of adding the genetic algorithm to the implemented algorithm, in which the lowest optimization percentage was approximately 1%, resulting in a portfolio balance of R$ 68,200.00. It is also noteworthy that on the occasions when situations of loss were observed, profits started to be made. It is worth noting that performance was improved not only by the portfolio result, but also by the reduction in risks, since fewer operations were carried out and with greater precision, which means, less exposure to risk.
The financial market has always been very attractive for the Brazilian population, however with the high interest rate it wasn’t worth the risk it offers in exchange for good financial returns. With the constant decrease in the base interest rate (Selic), Brazilian people started to seek alternatives to maintain an acceptable income and thus maintain or improve their financial condition. The number of CPFs registered at Brazilian’s stock market has been increasing exponentially in recent times, and this is due not only to the search for income alternatives, but also to digital influencers, who try to bring financial concepts in a simplified way in order to mature the vision of Brazilian population. One way to improve profitability is with the use of robot traders, which through chart analysis try to identify patterns, and thus obtain profit in operations based on decision making, identified by these patterns. To identify the potential of a trading robot assisted by artificial intelligence, initially a financial strategy was developed taking in account a risk management of the operations that will be executed by the robot. The defined strategy was implemented in the Metatrader 5 software, using the MQL5 programming language, which allows the performance of stock exchange operations in real time, as well as the tests of possible operations in the past, to simulate the behavior that can be achieved in the market. Once the code has been defined, it is now possible to apply optimization through the implementation of artificial intelligence, which should improve the risk management strategy, providing more accurate operations. The results obtained prove the efficiency of adding the genetic algorithm to the implemented algorithm, in which the lowest optimization percentage was approximately 1%, resulting in a portfolio balance of R$ 68,200.00. It is also noteworthy that on the occasions when situations of loss were observed, profits started to be made. It is worth noting that performance was improved not only by the portfolio result, but also by the reduction in risks, since fewer operations were carried out and with greater precision, which means, less exposure to risk.
Descrição
Palavras-chave
Mercado financeiro, Trader, Metatrader, Robô, Inteligência artificial, Financial market, Robot, Artificial intelligence