河北大学学报(自然科学版) ›› 2015, Vol. 35 ›› Issue (2): 182-187.DOI: 10.3969/j.issn.1000-1565.2015.02.013

• 电子工程与计算机科学 • 上一篇    下一篇

无监督排序学习算法的一致性比较分析

李纯果1,李海峰2   

  1. 1.河北大学 数学与信息科学学院,河北 保定 071002;2.河北大学 党委组织部,河北 保定 071002
  • 收稿日期:2014-09-12 修回日期:2014-11-12 出版日期:2015-03-25 发布日期:2015-03-25
  • 通讯作者: 李纯果
  • 作者简介:李纯果(1981-),女,河北邯郸人,河北大学讲师,主要从事模式识别理论基础与应用、机器学习等研究
  • 基金资助:
    河北省自然科学基金项目(F2013201060)

Comparison Analysis on Ranking Consensus

Li Chunguo1, Li Haifeng1   

  1. (1. College of Mathematics and Information Science, Hebei University, Baoding 071002, China; 2. Department of Party Committee Organization, Hebei University, Baoding 071002, China
  • Received:2014-09-12 Revised:2014-11-12 Online:2015-03-25 Published:2015-03-25
  • Contact: Li Chunguo

摘要: 对于无监督的排序学习算法来说,排序结果的评价指标是非常具有挑战性的问题.从一致性的角度,比较了4种比较典型的无监督排序学习方法,并在机器学习标准数据库中进行试验比较分析.结果显示,RPC这种非线性的无监督排序融合方法产生的排序结果有最小的Kendall距离和Spearman简捷距离,体现了RPC在无监督排序方法的优越性.

关键词: 启发式排序, 学习式排序, 排序一致性, RPC

Abstract: Unsupervised ranking has a big challenge that there is no standard metric to measure the ranking results. This paper tries to propose a comparison scheme for unsupervised ranking based on ranking consensus. Some metrics used in supervised ranking can be used here to measure unsupervised ranking consensus. However, these metrics are taken between ranking lists and attributes, rather than ranking lists and ranking labels. Experimental results on UCI datasets show that RPC, which is one kind of unsupervised ranking aggregation method, has the minimum Kendall Distance and Spearman Footrule Distance than the other representative unsupervised ranking methods.

中图分类号: