Prediction of stock market using systematic trading agent considering future values

Technical Analysis is a methodology that examines the value of market, which in turn is a reflection of all available market information and can be undervalued or overvalued. The analysis of historical asset value can indicate possible trends. Systematic Trading Agents are computer-based systems that employ historical prices as inputs to generate signals to enter or exit the market.

The objective of this work was to maximize monetary returns and proposes the following process: (1) multiple agents are tested and compared, resulting in better profits without taking into account the cost of executing several market transactions (Triple Exponential Moving Average gave better results in our experiments). (2) Future prices are predicted using recurrent neural network and fed to the agent, with the goal of improving the agent and reducing the number of spurious market transactions. The proposed model successfully outperforms multiple agents using data from the Spanish equity index IBEX-35.