Scholarlink

<p>Actionable knowledge discovery is an established discipline of data mining. The goal is to find patterns hidden in the data that suggest actions beneficial to the data owner. Action rules are tools of actionable knowledge discovery. They are based on flexible and stable attributes. Action rules suggest the change of flexible attributes, leading to the desirable reclassification of a target attribute. The change of flexible attributes can be applied to a group of objects described by stable attributes. We introduce histogram action rules that deal with histograms of target attributes. These action rules propose changes to values of flexible attributes that result in a new shape of the target attribute histogram, such that comparing the shapes before and after the change reveals valuable information.</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

Histograms as tools for mining actionable knowledge – principles and examples

  • 期刊论文

  • 发表日期：

    2025年1月24日

查看全文
  • Jan Rauch

摘要

Actionable knowledge discovery is an established discipline of data mining. The goal is to find patterns hidden in the data that suggest actions beneficial to the data owner. Action rules are tools of actionable knowledge discovery. They are based on flexible and stable attributes. Action rules suggest the change of flexible attributes, leading to the desirable reclassification of a target attribute. The change of flexible attributes can be applied to a group of objects described by stable attributes. We introduce histogram action rules that deal with histograms of target attributes. These action rules propose changes to values of flexible attributes that result in a new shape of the target attribute histogram, such that comparing the shapes before and after the change reveals valuable information.