Deep reinforcement learning for financial trading using multi-modal features

Loukia Avramelou, Paraskevi Nousi, Nikolaos Passalis, Anastasios Tefas

Abstract: Highlights•Deep Reinforcement Learning for financial trading using price and sentiment features.•Public opinion influences the cryptocurrency market and vice versa.•Price and sentiment information is beneficial in the field of cryptocurrency trading.•Multi-modal embedding layers can improve agents’ performance and explainability.