Portfolio Management Using Multi-Agent Reinforcement Learning

Data is one of the most crucial factors in the stock market that drives human decision-making. When properly analyzed, market data can provide market insights that prove extremely valuable in turning a profit for a client. Collecting and understanding market indicators can be extremely time-consuming and costly, whether a company or an individual. Automation of this process has been essential to the productivity and success of traders in this industry. In this presentation, I will give a walkthrough of the Deep Learning Reinforcement learning-based stock portfolio management system using multiple innovative indicators, e.g., news sentiments scores, and how these indicators and the environment enhanced the results.