Comparative analysis of forecasting portfolio returns using Soft Computing technologies

Abstract

We propose using fuzzy time series (FTS) to forecast the future performance of returns on portfolios. We model the portfolio selection problem by means of possibilistic moments, and approximate the uncertainty of the return on a given portfolio by trapezoidal fuzzy numbers.Some modifications into the classical models of FTS, based on weighted operators, allow us generating trapezoidal numbers as forecasters of the performance of portfolio returns. We incorporate our proposals into classical FTS methods and analyze their effectiveness with respect to FTS models with a possibilistic interval-valued mean approach, using historical returns on assets from the Spanish stock market.