Causal Bayesian Analysis for Cryptocurrency Reinforcement Learning

Summary

The paper introduces CausalReinforceNet, a framework intended to incorporate causal analysis into reinforcement-learning decisions for cryptocurrency trading. It combines Bayesian and dynamic Bayesian network techniques with reinforcement learning, aiming to help agents reason about market relationships when choosing trades. The study applies the framework to Binance Coin, Ethereum, Litecoin, Ripple, and Tether.

The authors build two agents using different reinforcement-learning algorithms and compare them with buy-and-hold and a baseline RL model. They report that the causal framework produces greater profitability than both comparison approaches, while its effectiveness differs across the cryptocurrencies studied. The summary does not state the sample period, transaction costs, risk-adjusted results, validation design, or whether the findings hold outside the evaluated assets. The reported profitability therefore describes the study’s comparisons rather than establishing general trading performance.

Key ideas

• CausalReinforceNet combines Bayesian and dynamic Bayesian network techniques with reinforcement learning.
• The framework is evaluated on five prominent cryptocurrencies.
• Two agents use different reinforcement-learning algorithms.
• The reported comparisons include buy-and-hold and a baseline reinforcement-learning model.
• Reported profitability varies among the cryptocurrencies, and the provided description omits key evaluation details.

Tags

# 2310.09462


# A Framework for Empowering Reinforcement Learning Agents with Causal Analysis: Enhancing Automated Cryptocurrency Trading









Despite advances in artificial intelligence-enhanced trading methods, developing a profitable automated trading system remains challenging in the rapidly evolving cryptocurrency market. This research focuses on developing a reinforcement learning (RL) framework to tackle the complexities of trading five prominent altcoins: Binance Coin, Ethereum, Litecoin, Ripple, and Tether. To this end, we present the CausalReinforceNet~(CRN) framework, which integrates both Bayesian and dynamic Bayesian network techniques to empower the RL agent in trade decision-making. We develop two agents using the framework based on distinct RL algorithms to analyse performance compared to the Buy-and-Hold benchmark strategy and a baseline RL model. The results indicate that our framework surpasses both models in profitability, highlighting CRN's consistent superiority, although the level of effectiveness varies across different cryptocurrencies.

Shown in full with attribution under the source's licence. Licence: abstract CC0

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.