What question did this study set out to answer?

Develop a high-frequency machine learning framework for short-horizon detection and monitoring of systemic financial stress across Europe.

August 18, 2026Decision Analytics JournalOpen Access

A machine learning framework for short-horizon monitoring of systemic financial stress in Europe

View Full Paper

Ask AI

Bookmark

Authors

DDDmytro Diachkov AAAfshin Ashofteh

Discussion

Member takes

Overview

Predictive modeling study demonstrates accurate short-horizon monitoring of European financial stress using volatility-based logit models, suggesting simpler tools excel in near-term surveillance.

Key Points

Develop a high-frequency machine learning framework for short-horizon detection and monitoring of systemic financial stress across Europe.
Transformed the European Central Bank Composite Indicator of Systemic Stress (CISS) into binary targets using fixed, rolling-sigma, and rolling-percentile threshold rules.
Evaluated linear and nonlinear classifiers across daily multi-asset datasets using chronological out-of-sample validation to eliminate look-ahead bias.
Assessed predictive accuracy across multiple forecast horizons using rare-event evaluation metrics, focusing on precision–recall AUC, recall, and F1 score.
Predictive performance peaked under a fixed CISS threshold at short forecast horizons, especially within a 5-business-day window.
A transparent logit model relying on CISS and VIX level and change features was the preferred specification, matching or exceeding more complex nonlinear models.
Rolling thresholds preserved qualitative short-horizon warning patterns but increased specification instability and reduced overall performance, especially at longer horizons.

Cite This Study

Diachkov et al. (2026) studied this question.

synapsesocial.com/papers/6a911c30c39ca0eda18310a0 https://doi.org/10.1016/j.dajour.2026.100742

View Full Paper

Ask AI

Bookmark

August 18, 2026Decision Analytics JournalOpen Access

A machine learning framework for short-horizon monitoring of systemic financial stress in Europe

View Full Paper

Ask AI

Bookmark

Authors

DDDmytro Diachkov AAAfshin Ashofteh

Discussion

Member takes

Overview

Predictive modeling study demonstrates accurate short-horizon monitoring of European financial stress using volatility-based logit models, suggesting simpler tools excel in near-term surveillance.

Key Points

Develop a high-frequency machine learning framework for short-horizon detection and monitoring of systemic financial stress across Europe.
Transformed the European Central Bank Composite Indicator of Systemic Stress (CISS) into binary targets using fixed, rolling-sigma, and rolling-percentile threshold rules.
Evaluated linear and nonlinear classifiers across daily multi-asset datasets using chronological out-of-sample validation to eliminate look-ahead bias.
Assessed predictive accuracy across multiple forecast horizons using rare-event evaluation metrics, focusing on precision–recall AUC, recall, and F1 score.
Predictive performance peaked under a fixed CISS threshold at short forecast horizons, especially within a 5-business-day window.
A transparent logit model relying on CISS and VIX level and change features was the preferred specification, matching or exceeding more complex nonlinear models.
Rolling thresholds preserved qualitative short-horizon warning patterns but increased specification instability and reduced overall performance, especially at longer horizons.

Cite This Study

Diachkov et al. (2026) studied this question.

synapsesocial.com/papers/6a911c30c39ca0eda18310a0 https://doi.org/10.1016/j.dajour.2026.100742

View Full Paper

Ask AI

Bookmark