International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 10 | Oct 2025
p-ISSN: 2395-0072
www.irjet.net
Investment Recommendation System using Reinforcement learning: A Survey Govardhan P. Bagul1, Prof. Pramila M. Chawan2 1M. Tech Student, Dept of Computer Engineering, and IT, VJTI College, Mumbai, Maharashtra, India
2Associate Professor, Dept of Computer Engineering, and IT, VJTI College, Mumbai, Maharashtra, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Investor choice is regularly hampered by
investigates how AI can help fill this gap by integrating behavioral finance knowledge into financial advisory systems.
behavioral biases including loss aversion, impulsiveness, and overconfidence, which are not catered to by conventional portfolio optimization models because of their rationality assumptions. With developments in Artificial Intelligence (AI) and Machine Learning (ML), it now becomes feasible to systematically identify and curb these biases so that more adaptive and tailored investment plans can be made. This paper surveys recent studies on AI technology in behavioral finance and suggests a multi-layered system architecture incorporating behavioral profiling, market data analysis, explainable AI, and continuous learning mechanisms. The proposed framework bridges available gaps in dynamic behavioral modeling and transparency and serves as a basis for intelligent, user-oriented financial advisory systems
2. LITERATURE SURVEY This survey discusses representative studies on the use of Artificial Intelligence (AI) in behavioral finance, sentiment analysis, and portfolio management. The chosen papers address: 1. The use of AI and behavioral economics to influence consumer financial health. 2. The detection of specific investor biases using machine learning. 3. The development of advanced deep reinforcement learning (DRL) frameworks for high-dimensional portfolio management. 4. The creation of a reinforcement learning framework that adapts to customizable stock pools without retraining 5. The application of sentiment analysis to predict market behavior.
Key Words: Behavioral Finance, Investor Biases, Artificial Intelligence, Machine Learning, Reinforcement Learning, Explainable AI, Portfolio Optimization, Adaptive Systems
1.INTRODUCTION Financial markets are conventionally described in the Efficient Market Hypothesis (EMH) framework, where it is assumed that investors are rational maximizers of utility. Behavioral finance refutes this postulate by showing that investors systematically depart from rationality because of cognitive and affective biases. Prospect Theory researches by Kahneman and Tversky (1979) cemented that people are disproportionately loss-averse relative to equivalent gains a phemonemon seen as loss aversion.
1. Ben-David, Mintz, and Sade, Working Paper (2020) AI and Behavioral Nudges for Overdraft Fee Reduction Ben-David et al. implemented a randomized field study with a large personal financial management platform to examine mechanisms for lowering overdraft charges. They employed an AI algorithm to classify users as likely to overdraft their accounts and dispatched them reminder alerts with varying framings. The research revealed that sending a reminder was effective, but its impact was greatly increased if the message was simplified. In addition, a negatively phrased simple message ("Avoid Paying $34.00 Fee") elicited more user activity and had a higher, longer-lasting effect on fee reductions than a positively phrased one ("Save $34.00"). The impacts were largest with users who had medium to high yearly incomes.
As digital trading platforms have become more prominent, retail investors create enormous volumes of transactional and behavioral data. Such data offer a chance to study impulsiveness, overconfidence, and herding patterns. At the same time, the financial sector is progressively embracing artificial intelligence for prediction, portfolio management, and risk evaluation. The problem lies in combining behavioral modeling with AI-based decision-making. The majority of AI investment systems maximize returns against market signals, not taking into consideration the behavioral aspect of the enduser. This causes a decoupling between model recommendations and the user's actual decision patterns, which can lower trust and adoption. This study
© 2025, IRJET
|
Impact Factor value: 8.315
2. Bhorge et al., IJIRT (2024) Financial Sentiment Analysis using Transformer Models Bhorge et al. surveyed recent advances in sentiment analysis for finance with an emphasis on sentiment extraction from
|
ISO 9001:2008 Certified Journal
|
Page 486