EUVIMEDEUVIMED entdecken
Crossref

Designing a Composite Comprehensive Stress Index for the Tehran Stock Exchange Using a Machine Learning Approach

Elham Babaheidarian, Farhad Hanifi, Mirfiz Fallah Shams

Accounting, Finance and Computational Intelligence · 2026 · 4 (1) · 1-28

DOI: 10.61838/jafci.207 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Accounting, Finance and Computational Intelligence, 2026. Verfasst von Elham Babaheidarian und 2 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This study aimed to design and validate a comprehensive index to monitor systemic risk and market-wide stress in the Tehran Stock Exchange using advanced econometric models and machine learning algorithms. Daily time-series data of selected Tehran Stock Exchange indices from 2014 to 2024 were analyzed. Logarithmic returns were calculated, and the DCC-MGARCH model was applied to estimate the dynamic conditional correlation matrix and systemic risk metrics such as ΔCoVaR. To determine feature importance and optimal weighting of indices, three supervised learning algorithms (support vector regression, artificial neural networks, and random forest) were compared, with random forest selected due to superior predictive accuracy. The composite stress index was then constructed and validated using time stability analysis, stress (shock) testing, and logistic regression forecasting. The results revealed that automobile, real estate, paper products, and metal products sectors carried the highest systemic risk, while computer, coal, and textiles showed the lowest. The comprehensive stress index provided reliable early warning signals during market turbulence and achieved strong predictive performance, with an AUC of 0.801 and an accuracy of 89.9% in logistic regression analysis for shock detection. The developed composite stress index is a robust and dynamic tool for identifying vulnerability points and forecasting systemic crises in the Tehran Stock Exchange. It offers significant practical value for policymakers, market analysts, and regulatory authorities to strengthen market resilience and implement proactive risk management strategies. Incorporating macroeconomic variables and extending the historical dataset could further enhance its accuracy and generalizability.

Zitieren

APA

Babaheidarian, E., Hanifi, F., & Shams, M. F. (2026). Designing a Composite Comprehensive Stress Index for the Tehran Stock Exchange Using a Machine Learning Approach. Accounting, Finance and Computational Intelligence, 4(1), 1-28. https://doi.org/10.61838/jafci.207

Vancouver

Babaheidarian E, Hanifi F, Shams MF. Designing a Composite Comprehensive Stress Index for the Tehran Stock Exchange Using a Machine Learning Approach. Accounting, Finance and Computational Intelligence. 2026;4(1):1-28. doi:10.61838/jafci.207

BibTeX
@article{Babaheidarian2026,
  author = {Babaheidarian, Elham and Hanifi, Farhad and Shams, Mirfiz Fallah},
  title = {Designing a Composite Comprehensive Stress Index for the Tehran Stock Exchange Using a Machine Learning Approach},
  journal = {Accounting, Finance and Computational Intelligence},
  year = {2026},
  volume = {4},
  number = {1},
  pages = {1--28},
  doi = {10.61838/jafci.207}
}
Kostenlos starten →