Machine Learning and Multi-criteria Analysis on the Forex Market

The forex market is considered the most liquid market in the world. Many instruments—currency pairs and market indicators are used in technical analysis. The decision-maker observes many buy and sell signals generated by the market indicators for different instruments in the investing period in which he/she can make respective trading decisions. The efficiency of the signals is typically analyzed post-factum and concerns one efficiency measure. What’s more, it is not included in existing decision-support trading systems. In this paper, a new idea is proposed in which the efficiency analysis is included in the trading system using a machine learning mechanism and is made using not one but some number of different efficiency measures. A respective multi-criteria optimization problem is formulated to select the Pareto-optimal variants of decisions that can be proposed to the decision-maker. This idea has been implemented in the proposed experimental computer-based system, including the algorithm deriving these Pareto-optimal variants.