Deep Learning vs. Reinforcement Learning in Automated Stock Trading: A PRISMA-Oriented Systematic Review

Wasim Qureshi; Arup Kadia; Ankush Goyal; Kamal Narayan

Deep Learning vs. Reinforcement Learning in Automated Stock Trading: A PRISMA-Oriented Systematic Review

Author(s)	Wasim Qureshi, Arup Kadia, Dr. Ankush Goyal, Kamal Narayan
Country	India
Abstract	The rapid development of artificial intelligence has significantly changed stock trading systems, mainly through its two major systems Deep Learning and Reinforcement Learning, which are the leading technologies in this area. This paper systematically reviews the literature on the basis of PRISMA guidelines to assess and compare the effectiveness of DL and RL automation systems in stock market trading. The authors retrieved journal articles from Scopus and Web of Science that they published between 2015 and 2025 according to their pre-designed inclusion and exclusion criteria. The team employed strict selection techniques to pick a few studies whose data were then subjected to qualitative and quantitative analyses. The review article essentially judges the models by the performance metrics profitability, risk, adjusted returns, drawdown control, transaction cost sensitivity, and the ability to cope with various market conditions. The research shows DL models which contain recurrent and convolutional architectures can predict market outcomes accurately during periods of market stability and trends, while RL-based systems can adapt to changing conditions and make dynamic choices and function effectively during periods of market instability. Systems that combine DL for feature extraction with RL for policy optimization achieve better results than systems that operate independently, which shows that adaptive learning systems are gaining more importance. The field has gone far in its development, but it still struggles with issues such as biased data and overfitting, along with very high computational requirements and problems in replicating results. This review first looks at the current state of research and, at the same time, points out the gaps in the research that exist in the AI, based trading systems that are suggested to run at scale without sacrificing risk management and transparency. The paper is a great resource for both researchers and practitioners engaged in the design of automated trading systems using deep learning and reinforcement learning methods for achieving peak performance.
Keywords	Deep Learning, Reinforcement Learning, Automated Stock Trading, Algorithmic Trading Systems, Performance Evaluation, PRISMA Systematic Review
Field	Engineering
Published In	Volume 17, Issue 1, January-March 2026
Published On	2026-02-27

View / Download PDF File

doi

CrossRef DOI is assigned to each research paper published in our journal.

IJSAT DOI prefix is
10.71097/IJSAT

Downloads

Research Paper Format Copyright Permission Form and Undertaking Form Cover Page Vol 16 Isu 3 Cover Page Vol 16 Isu 2 Cover Page Vol 16 Isu 1

All research papers published on this website are licensed under Creative Commons Attribution-ShareAlike 4.0 International License, and all rights belong to their respective authors/researchers.

CC-BY-SA

About IJSAT Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us	Message on WhatsApp	+91-9687-182-185	editor@ijsat.org

International Journal on Science and Technology

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

Deep Learning vs. Reinforcement Learning in Automated Stock Trading: A PRISMA-Oriented Systematic Review

Share this