Scholarlink

<p>This article presents a data-driven model for identifying current risks in non-financial companies, supporting managerial and supervisory decision-making. The model integrates signals from the global news database GDELT and the decentralized social media platform Nostr into a structured risk register aligned with COSO ERM and ISO 31000 frameworks. Using standardized indicators of relevance and dynamics, it enables comparable monitoring of changes in the risk environment over time and across different sources. Empirical results from both sources indicate remarkable differences while complementing each other: news data provides a more stable, event-driven signal, while decentralized signals enable earlier risk detection. Forward-looking validation demonstrates that these indicators contain economically meaningful information, improving the prediction of future systematic risk beyond traditional approaches. The model does not replace existing methods of risk management, but supplements them with continuous, data-driven input into the decision-making processes of management and supervisory boards.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A model for high-frequency detection of current risks based on news analysis and decentralized social networks

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月20日

查看全文
  • Timotej Jagrič,
  • Maša Galun

摘要

This article presents a data-driven model for identifying current risks in non-financial companies, supporting managerial and supervisory decision-making. The model integrates signals from the global news database GDELT and the decentralized social media platform Nostr into a structured risk register aligned with COSO ERM and ISO 31000 frameworks. Using standardized indicators of relevance and dynamics, it enables comparable monitoring of changes in the risk environment over time and across different sources. Empirical results from both sources indicate remarkable differences while complementing each other: news data provides a more stable, event-driven signal, while decentralized signals enable earlier risk detection. Forward-looking validation demonstrates that these indicators contain economically meaningful information, improving the prediction of future systematic risk beyond traditional approaches. The model does not replace existing methods of risk management, but supplements them with continuous, data-driven input into the decision-making processes of management and supervisory boards.