Scholarlink

In this study we present a prediction vector model where a set of stock price time series is pairwise tested for Granger causality over an expanding causal interval and a softmax function is used to aggregate the causal interval results. We apply graph theory to the prediction vectors from the model to generate a Granger causality network. Our model is tested using a small sample consisting of daily closing prices of eight sectoral U.S. financial market instruments over a 15-year period. Our results show a Granger causality network with a Hamiltonian path and evidence of bivariate long memories among the financial market instruments.

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 Measure of Bivariate Long Memories in Financial Time Series with Applications to Granger Causality Networks

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Charles Mutigwe

摘要

In this study we present a prediction vector model where a set of stock price time series is pairwise tested for Granger causality over an expanding causal interval and a softmax function is used to aggregate the causal interval results. We apply graph theory to the prediction vectors from the model to generate a Granger causality network. Our model is tested using a small sample consisting of daily closing prices of eight sectoral U.S. financial market instruments over a 15-year period. Our results show a Granger causality network with a Hamiltonian path and evidence of bivariate long memories among the financial market instruments.