SAMS Subject Classification: 10, 26

The term financial market refers broadly to any marketplace in which the trading of securities occurs. These markets provide an opportunity for realising profitable returns, although, it is well known that humans are unable to trade efficiently in these markets due to inherent limitations of the human brain. This problem has led to the rise of algorithmic trading and the implementation of machine learning in pursuit of more consistent returns. Reinforcement learning (RL) is a class of machine learning algorithms that has recently achieved notable results when implemented in a trading environment. A framework for evaluating RL algorithmic trading strategies, tailored to the Forex market is presented in this poster. The framework is partitioned into different stages, most notably, a pre-processing phase of market data, an implementation phase of trading strategies in an RL context, and back-testing and optimisation phases of applying the trained RL models. The proposed framework is then utilised to establish an optimised algorithmic trading model which forms the basis of a flexible decision support system (DSS) capable of identifying a suitable algorithmic trading strategy for each of a specified set of market conditions.