<![CDATA[The high volatility of cryptocurrency assets, particularly Ethereum, poses a significant challenge for investors during market downturns. Traditional passive strategies often lead to substantial capital erosion in bearish conditions. This study explores the application of Deep Reinforcement Learning (DRL) through the Deep Q-Network (DQN) algorithm to develop an adaptive trading agent. By integrating technical indicators—Relative Strength Index (RSI), Simple Moving Average (SMA), and Moving Average Convergence Divergence (MACD)—the proposed model aims to optimize decision-making processes. Experimental results using historical data from 2020 to 2026 demonstrate that while the market experienced a significant decline of 19.55%, the DQN agent successfully maintained capital stability with a marginal deviation of only -0.54%. This finding suggests that the DQN-based approach offers superior risk mitigation and capital preservation capabilities compared to conventional buy-and-hold strategies in volatile financial environments.]]>
Copyrights © 2026
