An Attempt to Construct Automated Stock Trading Strategies Based on Reinforcement Learning

Abstract

In this study, we attempted to optimize stock trading strategies using a Deep Q-Network (DQN) model based on the Relative Strength Index (RSI). By utilizing historical data from Apple Inc., we evaluated the profitability across different RSI intervals. The results indicated that the performance of trading strategies varied according to the RSI range. This research evaluates the effectiveness of trading strategies using reinforcement learning and provides foundational insights for future studies.