What is financial machine learning?
Artificial Intelligence & Machine Learning
“Financial Machine Learning” by Bryan T. Kelly and Dacheng Xiu is an extensive survey that spans 150+ pages, aimed at mapping the intersection of machine learning and finance.
Kelly & Xiu attempt to capture and synthesize the burgeoning literature in this niche but rapidly expanding field, focusing on how machine learning tools are being integrated into financial market analysis.
Professor Xiu speaking at the MIT Ai & Quant Conference seated next to Exodus Point’s Head of Data Peter Cotton.
The paper/book makes several important contributions to the field of financial economics:
- Literature Synthesis: Kelly and Xiu provide a comprehensive overview of current research, highlighting key studies that have advanced our understanding of financial markets through machine learning.
- Research Directions: The authors suggest promising future research directions, including more complex model applications and the exploration of economic mechanisms and market equilibria using machine learning.
- Cross-Disciplinary Appeal: The survey is written to cater to both financial economists and machine learning experts, encouraging cross-disciplinary collaboration.
Main Topics Covered?
- Return Prediction: The authors discuss how ML has become applied to predict returns, a cornerstone of financial analytics.
- Risk and Portfolio Management: Though some topics like risk modeling are less covered, the paper acknowledges the importance of machine learning in enhancing risk assessment models and portfolio optimization.
- Derivatives Pricing: Early applications of neural networks in finance, specifically in options pricing and modeling implied volatility surfaces, become highlighted.
- Corporate Finance and Beyond: The survey also touches upon how machine learning is branching out into other areas of finance, including corporate finance, entrepreneurship, and real estate.
The paper does an outstanding job of delineating the current landscape of machine learning applications in financial studies.
Not only outlines what has become accomplished but also points out the gaps and opportunities for further research. However, one limitation is the paper’s selective focus on certain topics while omitting others like detailed risk modeling, which could be crucial for practitioners in the field.
Kelly & Xiu advocate for a broader application of machine learning beyond just asset pricing, highlighting its potential in areas such as corporate governance and venture capital. This approach not only broadens the scope of machine learning in finance but also enriches the discussion about its practical implications.
Future Implications
Kelly and Xiu‘s survey indicates that the future of market research will likely see deeper integrations of machine learning, not just for enhancing analytical methodologies but also for understanding complex economic behaviors and market dynamics. This shift could fundamentally alter how financial markets are studied and understood.
Read the full paper/book:
Bryan T. Kelly, Dacheng Xiu :: SSRN
