Scholarlink

<p>The collapse of major cryptocurrency exchanges such as FTX has underscored the urgent need for effective systemic risk monitoring in blockchain-based financial markets. This study proposes a comprehensive Value-at-Risk (VaR)-based framework for assessing the risk of crypto assets and introduces a real-time early warning system that incorporates clustering algorithms to detect volatility shifts and concept drift. Using historical price data from seven major cryptocurrencies, we apply GARCH-family models with skewed-t distributions to estimate downside risk and validate the effectiveness of VaR under extreme market conditions. We then integrate unsupervised machine learning techniques, including K-means, hierarchical clustering, and DBSCAN, to classify evolving risk patterns and generate timely alerts. Experimental results confirm the robustness of the proposed framework in capturing tail risks and identifying systemic vulnerabilities. The findings offer practical implications for investors, trading platforms, and regulators seeking to manage emerging risks in decentralized financial ecosystems.</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

Risk monitor system for blockchain-based security issuing system

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年6月20日

查看全文
  • Ji Liu,
  • Zao Zhang,
  • Zheng Xu,
  • Yanli Li,
  • Dong Yuan,
  • Shiping Chen

摘要

The collapse of major cryptocurrency exchanges such as FTX has underscored the urgent need for effective systemic risk monitoring in blockchain-based financial markets. This study proposes a comprehensive Value-at-Risk (VaR)-based framework for assessing the risk of crypto assets and introduces a real-time early warning system that incorporates clustering algorithms to detect volatility shifts and concept drift. Using historical price data from seven major cryptocurrencies, we apply GARCH-family models with skewed-t distributions to estimate downside risk and validate the effectiveness of VaR under extreme market conditions. We then integrate unsupervised machine learning techniques, including K-means, hierarchical clustering, and DBSCAN, to classify evolving risk patterns and generate timely alerts. Experimental results confirm the robustness of the proposed framework in capturing tail risks and identifying systemic vulnerabilities. The findings offer practical implications for investors, trading platforms, and regulators seeking to manage emerging risks in decentralized financial ecosystems.