Desenvolvimento de software multi estratégia para negociar em bolsa de valores

Abstract

In order to collaborate with the agents participating in the capital market, the savers of resources, this work proposes the development of a software trader that makes it possible to make quotation predictions of actions, as well as operating based on the results obtained from the approximations when using the polynomial interpolation method Newton’s form as an indicator of buying or selling shares. making it possible to increase profits and reduce losses in investments. The shares of publicly traded companies were evaluated using the MetaTrader 5 platform as an auxiliary tool to perform data collection, as well as allowing the execution of software trader on the platform. The results are presented for alpha, beta, newtonian strategies with and without adjustment, where it can be seen that beta obtained superiority in the financial result, however, Newtonian without adjustment achieved the highest gross profit among the techniques, alpha presented very few negotiations and benefits low financial costs compared to the others. For Newtonians with and without adjustment, the reduction in the amount of trades carried out when applying the automatic adjustment is visualized. The work shows the possibility to carry out automated trading in the capital market, although the proposal did not initially solve the short-term investors’ problems, there is a generalist software trader that allows reuse. However, the objective of developing and comparing strategies for the software traderwas achieved. Future work can be developed from the proposed work, implementing artificial neural networks and new strategies.