Modelo Híbrido GJR-GARCH Nebuloso para a Previsão da Volatilidade em Mercados de Ações

Forecasting stock market returns volatility is a challenging task thathas attracted the attention of market practitioners, regulators andacademics in recent years. This paper proposes a Fuzzy GJR-GARCH model toforecast the volatility of S&P 500 and Ibovespa indexes. The modelcomprises both the concept of fuzzy inference systems and GJR-GARCH modelingapproach in order to consider the principles of time-varying volatility,leverage effects and volatility clustering, in which changes are catalogedby similarity. Moreover, a differential evolution (DE) algorithm issuggested to solve the problem of Fuzzy GJR-GARCH parameters estimation. Theresults indicate that the proposed method offers significant improvements involatility forecasting performance in comparison with GARCH-type models andwith a current Fuzzy-GARCH model reported in the literature. Furthermore,the DE-based algorithm aims to achieve an optimal solution with a rapidconvergence rate.