A neural-fuzzy system for forecasting
Zuohong Pan, Xiaodi Liu, O. Mejabi · 2002
This study introduces a neural-fuzzy system for financial modeling and forecasting. The new system combines a neural network with fuzzy logic, in which fuzzy rules replace the traditional crisp logic in the reasoning. The system is used to exploit financial market inefficiencies and extract nonlinear patterns. When used in forecasting S&P 500 index, the model's performance is compared with a random walk model, an ARIMA model and other more sophisticated econometric models, (e.g. ARCH model). The power and predictive ability of the models are evaluated on the basis of mean absolute error, root mean squared error, turning point prediction, pattern recognition, and the conditional efficiency in the sense of (Granger and Newbold, 1973) and (Fair and Shiller, 1990). The study showed a promising result for the neural-fuzzy system.