Multiple Stock Trading using Ensemble Strategy and Deep Reinforcement Learning

Abstract

The aim of this study is to create an automated trading system for trading just one stock. The use of Markov Decision Process to simulate the stock trading process (MDP). Framed as a matter of maximizing, the trading objective is expressed. Many studies are being conducted daily throughout the globe to accurately forecast share prices and assist all share market participants. The artificial neural network is mostly utilized by the prediction model. There are other additional algorithms that are used to predict stock prices. The DRL (Deep Reinforcement Learning) algorithm and reinforcement learning will be utilized in this research paper to forecast stock price and improve prediction accuracy. There are also some challenges which occur such as Data quality, Data availability and Overfitting. The major objective of this research study is to anticipate the price of stocks using artificial intelligence and algorithms like DRL (Deep Reinforcement Learning) and Reinforcement Learning. For this case study, the data from a single stock that was pulled through the Yahoo Finance API (Application Programming Interface) will be used. The data includes Open-High-Low-Close price and volume.