Scholarlink

<p>The conditional interaction information of m discrete random variables (factors), given a discrete variable, is statistically estimated via the classical delta method for any positive integer m. The limiting distribution of the constructed estimates is found when the factors are conditionally independent for a given extra one. Previously, the case of m=2 was considered in the literature. The criterion of conditional independence of a set of random variables for a given random variable is established.</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

Statistical Estimation of Conditional Interaction Information by Delta Method

  • 期刊论文

  • 发表日期：

    2025年7月22日

查看全文
  • Alexander Bulinski,
  • Shanwen Wang

摘要

The conditional interaction information of m discrete random variables (factors), given a discrete variable, is statistically estimated via the classical delta method for any positive integer m. The limiting distribution of the constructed estimates is found when the factors are conditionally independent for a given extra one. Previously, the case of m=2 was considered in the literature. The criterion of conditional independence of a set of random variables for a given random variable is established.