Quantitative Analysis of Herding Effect Generation Mechanisms and Transmission Pathways in Financial Markets

Abstract

Herding effect is a typical irrational behavior in financial markets, and conducting quantitative analysis on it is of great significance for understanding abnormal market fluctuations. By constructing multi-dimensional emotional contagion dynamics models and complex network transmission frameworks, we can deeply analyze the generation mechanisms and diffusion pathways of herding behavior. Using high-frequency trading data to validate theoretical models, the research finds that for every 0.1 increase in network density, the transmission speed of herding effects increases by approximately 23.7%. Empirical analysis shows that during the 2015 stock market anomalous fluctuations, the herding effect transmission speed reached 15.3% per hour with an impact range covering 87.6% of the market. In machine learning prediction models, the LSTM algorithm achieved an accuracy of 87.3% and could identify herding behavior outbreak signals 2-3 trading days in advance.