多尺度决策系统中代价敏感的最优尺度组合
作者:
作者单位:

重庆邮电大学

作者简介:

通讯作者:

中图分类号:

TP18

基金项目:

国家自然科学基金项目(面上项目,重点项目,重大项目)(No.61876201);


Cost-sensitive optimal scale combination in multi-scale decision systems
Author:
Affiliation:

Chongqing University of Posts and Telecommunications

Fund Project:

The National Natural Science Foundation of China (General Program, Key Program, Major Research Plan)(No.61876201);

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    最优尺度组合是多尺度决策系统中研究的热点之一,现有的研究大多是从一致性、不确定性的角度出发,而没有充分考虑代价信息的影响.针对这个问题,本文首先分析了最优尺度组合中考虑代价敏感的重要性,从决策代价的角度提出基于测试代价和延迟代价的多尺度决策系统,并且定义了尺度代价和属性代价来刻画尺度和属性所产生的代价;其次,考虑实际场景中属性代价的影响,将属性重要度和属性代价结合进行属性排序;最后,属性进行最优尺度选择时,考虑尺度代价的影响,建立了一个代价敏感的最优尺度组合选择模型.实验结果表明,在现有代价认知场景下,该模型能合理地进行最优尺度组合选择,所得结果更符合实际需求.

    Abstract:

    Optimal scale combination is one of the research hotspots in multi-scale decision systems. Most existing researches start from the perspective of consistency and uncertainty, without fully considering the influence of cost information. In view of this issue, the importance of considering cost-sensitivity in optimal scale combination is analyzed in this paper firstly, a multi-scale decision system based on test cost and delay cost from the perspective of decision cost is proposed, and the scale cost and attribute cost to describe the cost caused by scale and attribute are defined. Secondly, considering the effect of attribute cost in the real scene, attribute significance and attribute cost are combined to sort the attributes. Finally, a cost-sensitive optimal scale combination selection model is established by considering the influence of scale cost when selecting the optimal scale of attributes. The experimental results show that with the existing cost cognition scenario, the model proposed can reasonably select the optimal scale combination, which is more in line with the actual requirements.

    参考文献
    相似文献
    引证文献
引用本文
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2020-02-08
  • 最后修改日期:2020-04-14
  • 录用日期:2020-04-22
  • 在线发布日期:
  • 出版日期: