Systemic Risk Contagion and Intelligent Early Warning: A Dynamic Network and Machine Learning Approach
Advances in Engineering Technology Research · 10.03.2026 · 16 (1) · 462
DOI: 10.56028/aetr.16.1.462.2026 ↗Kurz gefasst
Fachartikel in der Zeitschrift Advances in Engineering Technology Research, online erschienen am 10.03.2026. Verfasst von Kaining Wang.
Abstract (Originalsprache)
This paper contributes to the development of an innovative multi-stage system for contagion modeling for financial markets. To begin, we use a DCC GARCH model to analyze the changing relationships of a large collection of financial institutions. To derive a non-dense and meaningful network representation from a dense correlation matrix, we make specific use of the Graphical LASSO (GLASSO) method. With this approach, we arrive at sparse and thus meaningful realizations of dynamic precision matrices that depict conditional dependencies. With the use of an advanced approach, we are able to measure the extent and direction of risk spillovers and form the Diebold-Yilmaz spillover index. Thus, a real-time, detailed risk transmission assessment and identification of major sources of institutional super-spreading is created. Since we have obtained a large set of features via dynamic network topology and spillover analysis, the last step is to use them in the LSTM neural network. The purpose of the network is to predict systemic events, which refers to drawing down or rallying in the market as captured in financial stress indices. Our hypothetical empirical analysis of the US financial sector indicated that this heterogeneous framework has a significant out-of-sample forecasting improvement over the others, as indicated by benchmark models such as Logistic regression. Moreover, we utilize SHAP (Shapley Additive exPlanations) in our black-box EWS to create an additional layer of logic by identifying the risk factors.
Zitieren
APA
Wang, K. (2026). Systemic Risk Contagion and Intelligent Early Warning: A Dynamic Network and Machine Learning Approach. Advances in Engineering Technology Research, 16(1), 462. https://doi.org/10.56028/aetr.16.1.462.2026
Vancouver
Wang K. Systemic Risk Contagion and Intelligent Early Warning: A Dynamic Network and Machine Learning Approach. Advances in Engineering Technology Research. 2026;16(1):462. doi:10.56028/aetr.16.1.462.2026
BibTeX
@article{Wang2026,
author = {Wang, Kaining},
title = {Systemic Risk Contagion and Intelligent Early Warning: A Dynamic Network and Machine Learning Approach},
journal = {Advances in Engineering Technology Research},
year = {2026},
volume = {16},
number = {1},
pages = {462},
doi = {10.56028/aetr.16.1.462.2026}
}
EUVIMED