Stock Market Investment Strategy Using Deep-Q-Learning Network

Abstract

Artificial intelligence demonstrates its ability to analyze time series data more efficiently than humans and to automate stock trading processes without the need for human interaction. Developing a stock market investment strategy using artificial intelligence (AI) involves leveraging AI techniques to analyze data, identify patterns, and make informed investment decisions. In this study, we commonly utilized reinforcement learning DQN to create algorithms. Deep Q-Learning Network (DQN) is a reinforcement learning algorithm that combines Q-learning with deep neural networks to handle high-dimensional state spaces. It was introduced by Deep Mind in 2013 and has since been applied to various domains, including gaming, robotics, and finance. Financial markets are highly complex and subject to various external factors, making the application of DQN in stock market investment is a challenging. A deep q-learning network maps the agent’s actions to its states. In addition, we performed external prediction in order to use forecasted prices as new features, changed the exploration rate, and implemented a stop-loss technique in order to increase trading performance. A list of the most predictable stocks is also supplied as an optional and referable rank list, which is a marginal output of external prediction. The final AI trader has the capacity to study from previous years’ data of three selected equities and trade automatically with significant earning potential using new data. Our AI trading system model that uses reinforcement learning and deep learning on time series to trade stocks in the stock markets intelligently and safely, with estimated returns of 0.425% per five trading days. Overall our investment strategy model achieved 23.4% return for 275 trading days. Developed model definitely helpful to Investment Bankers, mutual fund managers, and Individual Investors.