Hybrid flexible neural tree for exchange rates forecasting

Abstract

Exchange rate is an important link of international economic relations. In this paper, a novel method for improving flexible neural tree is proposed to forecasting exchange rate data. The hybrid flexible neural tree with pre-defined instruction sets can be created and evolved. The structure and parameters of hybrid flexible neural tree is optimized using probabilistic incremental program evolution and particle swarm optimization algorithm. Compared with the conventional artificial neural network and flexible neural tree based on gene expression programming, the experimental results indicate that the proposed method is feasible and efficient.