This study aims to evaluate how deep reinforcement learning (DRL) can improve financial portfolio management. It also has a second goal of understanding if financial fundamental features (e.g., revenue, debt, cash flow) improve model performance. After conducting a literature review to establish the current state-of-the-art, the CRISP-DM method was followed: 1) Business understanding; 2) Data understanding; 3) Data preparation on two datasets, one with market only features and another with also fundamental features; 4) Modeling – Advantage Actor-Critic, Deep Deterministic Policy Gradient and Twin-delayed DDPG DRL models were optimized; 5) Evaluation. Models had a consistent sharpe ratio performance across datasets – average of 0.35 vs 0.30 for the baseline, in the test set. It is also demonstrated that fundamental features improved model robustness and consistency. Hence, supporting the use of both DRL models and quantamental investment strategies for portfolio managers to generate alpha while increasing investor’s trust through higher transparency

--