Prospecção do uso de machine learning nas corretoras brasileiras

Abstract

Machine learning technology is increasingly used in the US, Europe and Asia financial sector. In this context, the present study sought to prospect data regarding the use of this technology by Brazilian securities brokers. Two hypotheses were investigated: the use of this technology or not by brokerages and, if so, in which they are applied. An analytical descriptive methodology was adopted. A two part onl ine form was made available to 72 brokerages in the country, with questions that define their profile and show the uses attributed to machine learning. The results were limited to the brokers’ low adherence to the questionnaire that only obtained four answ ers, representing 5% of the population. The study shows that machine learning is used by the brokerage firms investigated, but at a level below its potential. Its purpose includes analyzing financial data and predicting market trends. The use of ML was als o pointed as relevant to increase revenue and productivity of operations of companies that participated in the research. Thus, it is noted that the machine learning techniques are little explored for financial operations in Brazil. However, there are restrictions on these findings because of the limited sample size. Thus, it is necessary to expand studies on the subject to guide the country’s financial institutions in the adoption of this technology, as it would help in understanding the advantages and disadvantages of using it.