Reinforcement Learning for Stock Price Trading with Keywords in Google Trends

Abstract

In this paper, we apply the Proximal Policy optimization (PPO) algorithm to train an agent for automated stock trading. In additional to the conventional trading indicators, we also add the strength of keywords obtained from the Google Trends for training the agent. We conduct two experiments to test the effectiveness of adding keywords. The first experiment uses general keywords, such as inflation. The second experiment uses stock-specific keywords, such as AAPL for trading apple stock. The experimental results confirm that the proposed approach can improve trading performance.