Modelos de lógica y lógica borrosa en la predicción del IPSA

Abstract

The methodology of Fuzzy Logic, is based on the idea that variables must be handled not as numbers but rather by the characteristics that they display. A series of historical daily quotations of the “Índice de Precios Selectivo de Acciones” of the Chilean stock market, corresponding to the period between the 14 of July of 1997 and the 07 of January of 2005 was used. It was constructed a model of logic and another one of fuzzy logic, in order to forecast the sign of the variations of the stock-exchange index already indicated. The models of logic and fuzzy logic registered a statistically significant forecast capacity. In addition, both models obtained extranormal returns statistically significant and positive when they were used in to a trading strategy, transaction costs considered.