Scholarlink

<p>Extracting and summarizing all the relevant information of a database to explain a variable of interest is an essential objective in data science. In this article, we introduce the notion of characterization of a Bernoulli variable. We give a precise definition and present methods for constructing minimal characterizations. Characterization synthesizes information on the variable of interest in the database. Therefore, it makes it possible to summarize the database with respect to this variable of interest. To illustrate this, we perform experiments on some databases to find characterizations and deduce summaries of these databases with respect to some variables of interest.</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

Characterization, a new tool for summarizing databases

  • 期刊论文

  • 发表日期：

    2025年3月28日

查看全文
  • Abdelkader Mokkadem,
  • Mariane Pelletier,
  • Louis Raimbault

摘要

Extracting and summarizing all the relevant information of a database to explain a variable of interest is an essential objective in data science. In this article, we introduce the notion of characterization of a Bernoulli variable. We give a precise definition and present methods for constructing minimal characterizations. Characterization synthesizes information on the variable of interest in the database. Therefore, it makes it possible to summarize the database with respect to this variable of interest. To illustrate this, we perform experiments on some databases to find characterizations and deduce summaries of these databases with respect to some variables of interest.