Zequn Li, Ph.D. Candidate in Financial Engineering

Bio

Zequn Li is a Ph.D. candidate in financial engineering at Stevens Institute of Technology, specializing in empirical asset pricing and interpretable machine learning. Her research explores the dynamics between firm characteristics and stock returns, with an expanding interest in fintech and financial analytics. 

Skillsets

Zequn is proficient in Python, R, SQL and machine learning, applying these skills to empirical asset pricing and statistical analysis. She combines programming and analytical abilities with experience teaching courses like programming for finance, statistics, and stochastic calculus, emphasizing practical application to prepare students for careers in the financial industry. 

Dissertation Summary

Interpreting Machine Learning Model in Empirical Asset Pricing

This dissertation advances our understanding of how firm characteristics predicts cross-sectional expected stock returns by integrating traditional asset pricing models with advanced machine learning techniques. It focuses on enhancing the interpretability of machine learning models using methods like Local Interpretable Model-Agnostic Explanations (LIME) and attention mechanisms.

A significant contribution of this research is the development of a novel framework that interprets machine learning asset pricing models through the LIME method. This approach illuminates how the inclusion of LIME local coefficients, representing interactions among characteristics within machine learning models, modifies the relationship between firm characteristics and stock returns. The empirical results underscore the importance of incorporating moderation effects into portfolio analysis. Specifically, certain firm characteristics exhibit varying long-short portfolio performance across different LIME groups, suggesting their predictive power is specific to certain asset segments. These findings deepen our understanding of the complexities in cross-sectional stock returns, uncovering detailed dynamics between firm characteristics and their return effects, and distinguishing this research from existing studies.

Furthermore, the dissertation introduces an innovative framework to interpret the behaviors of firm characteristics in predicting expected returns through machine learning models, directly addressing challenges of transparency and interpretability. By utilizing LIME, firm characteristics are evaluated based on their statistical significance and behaviors—linearity, independence, insignificance, and interaction—offering a novel perspective on their predictive roles. Empirical findings demonstrate a complex interplay among these behaviors, with interaction effects playing a pivotal role. This challenges the traditional emphasis on linear and independent influences in asset pricing models. The research provides new insights into the mechanisms of machine learning predictions in asset pricing, paving the way for further exploration into the economic rationale behind data-driven findings and enhancing the understanding of complex asset pricing dynamics.

Additionally, the dissertation incorporates attention mechanisms within the theoretical asset pricing framework. The attention mechanism allows the model to focus on the most relevant firm characteristics, improving both predictive performance and interpretability. By highlighting the most influential features, the attention mechanism enhances the model’s ability to predict expected returns more accurately and offers a clearer understanding of how specific characteristics influence stock returns.

Overall, this dissertation makes significant contributions to the literature on empirical asset pricing by bridging financial theories with machine learning techniques. By integrating interpretable machine learning methods like LIME and incorporating attention mechanisms, the research not only addresses challenges of complexity and interpretability in machine learning models but also uncovers detailed dynamics between firm characteristics and their effects on stock returns. This work offers a novel perspective on how firm characteristics influence future stock returns, distinguishing it from existing studies and paving the way for future research in financial engineering and analytics.