A fuzzy multicriteria approach for the trading systems on the Forex Market

Abstract

The paper relates to the trading systems supporting traders making decision on the forex market. Typical trading systems using technical analysis generate a buy or sell signal when the technical indicator crosses a given oversell or overbought levels. The paper extends the approach in which the above strict crisp conditions are replaced by fuzzy relations. The indicators are treated not independently as it is in the typical systems but jointly. Currency pairs are compared in the muliticriteria space in which each criterion is defined by a membership function referring to a given indicator. New formulations of the membership functions for different indicators are proposed. General ideas of the algorithm generating non-dominated alternatives in the multicriteria space are presented. The algorithm has been implemented in an experimental system. Computational results for different time windows using real-world data from the forex market are presented and discussed.