Modelagem, Implementação e Avaliação de Estratégias de Negociação Baseadas em Algoritmos de Aprendizado de Máquina para o Mercado Financeiro

Abstract

Investing in the stock market is one of the fastest and most attractive ways to make considerable prots in a short period of time. However, due to large variations and uctuations in this type of market, investors are subject to risks that can also result in large losses. In order to avoid that other students and interested in the nancial market area have to spend a lot of time on their research in the implementation of algorithms and can dedicate eorts in creating, validating and improving their trading strategies, this work proposes the design and implementation of a automated framework consisting of 5 stages: Data Extraction, Data Characterization and Transformation, Classication of Trend Forecasting Models, Operation Strategy and Results Analysis. During the simulations, historical quotation data of 9 assets traded on the Brazil Balcão Exchange (B3) was evaluated, for a period of 741 in the Validation stage, for the 8 proposed trend forecasting models. As a result and validation of the proposed framework, a consolidated table containing data (performance, operation / risk and statistics) and 2 graphs: series of closing price and series of accumulated capital (liquid and gross returns and operating cost) evolution of trends will be presented for each of the assets and trend forecasting models