Investor Sentiment and Market Interaction Modeling Irrationality Detection

Author: Anqi Liu

Date: April 26, 2017
Department: Financial Engineering
Degree: Doctor of Philosophy

Advisory Committee:
Dr. Steve Yang, Chairman
Dr. Zhengyu Cui,
Dr. Babak Heydar,
Dr. Khaldoun Khashanah,
Dr. Mark E. Paddrik

Abstract: This dissertation investigates how information and investor sentiment shape financial market dynamics through a behavioral finance perspective. With the rise of digitization and advanced data mining, sources such as business news and social media increasingly influence trading and investment decisions. We first study how sentiment propagates across investors using an agent-based framework and show that information sharing fosters the formation of common social beliefs about markets. We then examine the transmission of shocks in prices and news by modeling their arrivals with Multivariate Hawkes processes, documenting cross-excitation effects between market returns and investor sentiment. These findings reveal a feedback loop where price jumps alter social beliefs, and investor sentiment in turn influences trading and market movements. Finally, we propose an entropy-based method to detect irrational trading behavior, distinguishing between news-driven and price-driven activities. Overall, this work presents a behavioral perspective on financial markets under rational and irrational regimes and highlights the financial system as an evolving ecosystem shaped by adaptive investor behaviors.

For full Dissertation, click here.