Scholarlink

This paper introduces a novel hidden Markov graphical model to explore temporal interconnections among diverse financial markets. Exploiting the properties of the generalized hyperbolic family of distributions with time-varying parameters governed by a latent Markov chain, we can establish conditional correlation structures among variables. Our approach employs a penalized EM algorithm with an \(L_1\) penalty of the off-diagonal elements of the precision matrix. We apply the model to daily returns of metal, energy commodities and green energy ETFs from 2017 to 2024, showcasing its practical utility in unveiling time-varying network dynamics.

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

Hidden Markov Graphical Models with Generalized Hyperbolic Distributions: A Financial Analysis on Commodities and Green Energy Indexes

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Beatrice Foroni,
  • Luca Merlo,
  • Lea Petrella

摘要

This paper introduces a novel hidden Markov graphical model to explore temporal interconnections among diverse financial markets. Exploiting the properties of the generalized hyperbolic family of distributions with time-varying parameters governed by a latent Markov chain, we can establish conditional correlation structures among variables. Our approach employs a penalized EM algorithm with an \(L_1\) penalty of the off-diagonal elements of the precision matrix. We apply the model to daily returns of metal, energy commodities and green energy ETFs from 2017 to 2024, showcasing its practical utility in unveiling time-varying network dynamics.