Economic Theory and Machine Learning Integration in Asset Pricing and Portfolio Optimization: A Bibliometric Analysis and Conceptual Framework

Abstract

The integration of Machine Learning (ML) with economic theory has transformed financial market analysis, particularly in asset pricing and portfolio optimization. This study synthesizes existing research, identifies gaps and elucidates ML's impact on enhancing economic models and strategies through a Systematic Literature Review (SLR) of 401 relevant documents using bibliometric methodology and VOS viewer. Results show an increasing trend in publications, with the United States and China leading and institutions like the Shanghai University of Finance and Economics and Massachusetts Institute of Technology playing crucial roles. Cluster analysis reveals five main themes: Asset pricing and predictive analysis, algorithmic trading, data-driven portfolio management and optimization, reinforcement learning and adaptive strategies in financial markets and cryptocurrency market predictions. The proposed conceptual framework encompasses data acquisition and management, preprocessing and feature engineering, model selection and training, constraint formulation, theory-driven validation and real-time adaptation and monitoring, highlighting the potential synergy between ML and economic theory. This study provides insights into ML and economic theory integration, offering a structured pathway for enhancing asset pricing models and portfolio optimization techniques. Future research should refine this integration, focusing on practical application and adaptability to new financial instruments and market conditions