Intelligent computing and foreign exchange rate prediction : what we know and we don't

While exchange rate prediction has come a long way since the developments of early macroeconomic models and has embraced a number of intelligent computing (i.e., computational intelligence) methodologies in recent undertakings, there remain a number of uncertainties and debates around the characteristics of these different models. Further, while finance community is primarily focused on linear models due to interpretability, computational intelligence community has mainly focused on the time series aspects –leading to a gap between the two communities. This article provides a contemporary survey on the different exchange rate models from both the finance and computational intelligence domains. As illustrated, while computational intelligence has considerably progressed in this area, there still remains a number of research issues before these models will be well-accepted by the finance community. Thus, this article serves both as a comprehensive survey and link to future research opportunities for the domain.