Analysis of integration of financial series classification and constrained portfolio optimization

This paper proposes a new way of combining classification of financial asset price series and portfolio optimization, in which the optimization model has a constraint that ensures that the best rated assets in the classification process are present in the portfolio. This proposed combination is compared to each of the approaches applied separately. Trading simulation in the stock market using each of the strategies is carried out to evaluate the performance of each strategy, considering monthly performance metrics. Results show that portfolio optimization provides a lower risk, but classification shows great potential for a rapid increase in financial gain. The proposed combination manages to provide a greater gain than portfolio optimization, with a statistically equal risk, therefore, surpassing each of the individual approaches.