An automated econometric decision support system: forecasts for foreign exchange trades

Making decisions challenges foreign exchange (FX) market brokers due to the volatility of the foreign exchange market, as well as the unmanageable flood of possibly relevant information. Thus, decision making in this complex and dynamically changing environment is a difficult task requiring automated decision support systems. In this contribution, we describe an econometric decision support approach, which enables the extraction of essential information indispensable to set up accurate forecasting models. Our approach is based on a genetic algorithm (GA) and applies the resulting models to forecast daily EUR/USD-exchange rates. In doing so, the genetic algorithm optimizes single-equation regression forecast models. The approach discussed is new in literature and, moreover, allows flexibility in automated model selection within a reasonably short time.