Artificial intelligence (AI) models can help investors find portfolios in which the focus is to optimize the risk-return relationship. There are several algorithms and techniques in the literature that allow the application of tests to a set of historical data for the selection and validation of investment portfolios. Based on this, this research intends to examine the contribution of the main machine learning techniques used in portfolio management through a systematic literature review. By using the Methodi Ordinatio for selection and ranking of articles, we classified papers considering object of study, type of AI used, period of analysis, data frequency, balance and cardinality. In addition, we detail the main contributions and trends conceived until the year 2020. Therefore, our findings reveal gaps and suggest future works on the topic.

Ethanol is one of the most used fuels in Brazil, which is the second-largest producer of this biofuel in the world. The uncertainty of price direction in the future increases the risk for agents operating in this market and can affect a dependent price chain, such as food and gasoline. This paper uses the architecture of recurrent neural networks—Long short-term memory (LSTM)—to predict Brazilian ethanol spot prices for three horizon-times (12, 6 and 3 months ahead). The proposed model is compared to three benchmark algorithms: Random Forest, SVM Linear and RBF. We evaluate statistical measures such as MSE (Mean Squared Error), MAPE (Mean Absolute Percentage Error), and accuracy to assess the algorithm robustness. Our findings suggest LSTM outperforms the other techniques in regression, considering both MSE and MAPE but SVM Linear is better to identify price trends. Concerning predictions per se, all errors increase during the pandemic period, reinforcing the challenge to identify patterns in crisis scenarios.

Este estudo investiga estratégias de negociação em Fundos de Investimento Imobiliários (FIIs) utilizando Reinforcement Learning (RL) para maximizar o retorno de uma carteira. Foram implementados cinco algoritmos de RL, configurados com o método ator-crítico, para gerar estratégias de negociação, que foram avaliadas e comparadas com métodos tradicionais, como Buy and Hold (BH), mínima variância e o Índice IFIX. A amostra final, composta por 26 fundos, foi selecionada considerando critérios como histórico mínimo de 5 anos e volume de negociação. As estratégias de RL superaram a mínima variância e tiveram desempenho comparável ao BH, mas não superaram o IFIX. Isso sugere que o mercado de FIIs pode ser eficiente, limitando a eficácia de técnicas avançadas como RL e dificultando a superação do índice. As contribuições do estudo incluem demonstrar o potencial do RL em superar estratégias tradicionais em determinados contextos, oferecendo possíveis compreensões para investidores, gestores de carteiras e literatura científica.