• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>International Conference on Inventive Computation Technologies >A method for discrimination prevention in automated decesion making
【24h】

A method for discrimination prevention in automated decesion making

机译：一种自动决策中防止歧视的方法

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Data mining is popular technology for extracting and analyzing large amount of data. Data mining plays important role in automated decision making. The automated decision making process may give a sense of fair decision but, there can be discrimination as training data may act as source of discrimination. Discrimination can be direct and indirect. Method described in this paper modifies original data in order to prevent discrimination that results into extraction of discrimination free rules after mining process. This approach consists of discrimination discovery and data transformation as major step. It deals with implementation of direct rule protection algorithm and aims at analyzing the effect of extended lift. The qualitative analysis of the modified data is done by extracting classification rules from modified data and comparing its value of discrimination measure with threshold value to ensure removal of discrimination from the original data.
机译：数据挖掘是用于提取和分析大量数据的流行技术。数据挖掘在自动化决策中起着重要作用。自动化的决策过程可能会给人以公平的决策感，但是由于培训数据可能会成为歧视的来源，因此可能会有歧视。歧视可以是直接的，也可以是间接的。本文所述的方法会修改原始数据，以防止在挖掘过程后导致提取无歧视规则的歧视。这种方法包括判别发现和数据转换作为主要步骤。它涉及直接规则保护算法的实现，旨在分析扩展提升的效果。对修改后的数据进行定性分析是通过从修改后的数据中提取分类规则，并将其判别度量值与阈值进行比较，以确保从原始数据中消除判别。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号