News Release

AI and network science for emotional risk measurement in stock markets

Peer-Reviewed Publication

KeAi Communications Co., Ltd.

Fig. 7. Distribution of emotional vulnerability and network structural characteristics

image: 

Fig. 7 visualizes the relationship between the emotional vulnerability of CSI 300 constituent stocks and network structural characteristics. In this study, emotional vulnerability is divided into five ascending levels (Group 0 to Group 4) corresponding to low vulnerability (low curvature) to high vulnerability (high curvature), marked by points of different colors.

view more 

Credit: Chen, J., Chen, Z., Zhang, N., & Zheng, Y.

Financial markets are increasingly shaped by investor emotions, yet most sentiment indicators focus only on whether investors feel optimistic or pessimistic. A new study published in Risk Sciences introduces a different perspective: instead of sentiment direction, it measures how fragile or unstable investor emotions are.

Researchers from the Central University of Finance and Economics in Beijing propose a novel concept called emotional vulnerability, which captures fluctuations and instability in market sentiment. The study combines large language models (LLMs) with network structure analysis to quantify this previously unexplored dimension of investor behavior.

The team analyzed millions of stock forum posts from China’s A-share market. Using the Moka Massive Mixed Embedding (M3E) model, an advanced large language model for Chinese text, they converted investor comments into numerical vectors. These vectors were then used to build sentiment networks for individual stocks, where connections reflect similarities in investor opinions.

“The vulnerability of each sentiment network was measured using Ricci curvature, a concept from geometry that has been widely applied in financial system risk analysis,” shares corresponding author Ning Zhang, a professor at the university. “In this context, higher curvature indicates tighter connections and more homogeneous investor sentiment, making markets more fragile when disrupted.”

The results showed that when emotional vulnerability is high, stock prices tend to fall or become unstable, whereas when emotional vulnerability is low, markets exhibit stronger resilience and generate higher excess returns.

“We also found that portfolios consisting of emotionally stable stocks consistently outperform those with high emotional vulnerability. This relationship remains significant even after controlling for traditional asset pricing factors, including the Fama–French five-factor model, as well as for sentiment direction (optimistic or pessimistic),” adds Zhang.

By shifting attention from sentiment direction to sentiment structure, the research expands the toolkit of empirical asset pricing. The authors suggest that emotional vulnerability can improve risk assessment, enhance market trend prediction, and support more informed investment decision-making.

###

Contact the author: Ning Zhang, Central University of Finance and Economics

CUFE Chinese Fintech Research Center Beijing 102206, China, zhang-ning@vip.163.com

The publisher KeAi was established by Elsevier and China Science Publishing & Media Ltd to unfold quality research globally. In 2013, our focus shifted to open access publishing. We now proudly publish more than 200 world-class, open access, English language journals, spanning all scientific disciplines. Many of these are titles we publish in partnership with prestigious societies and academic institutions, such as the National Natural Science Foundation of China (NSFC).

 


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.