南京大学学报(自然科学版) ›› 2022, Vol. 58 ›› Issue (1): 9–18.doi: 10.13232/j.cnki.jnju.2022.01.002

• • 上一篇    下一篇

多用户偏好下基于三支决策的动态属性约简

刘鑫, 胡军(), 张清华, 于洪   

  1. 计算智能重庆市重点实验室,重庆邮电大学,重庆,400065
  • 收稿日期:2021-06-23 出版日期:2022-01-30 发布日期:2022-02-22
  • 通讯作者: 胡军 E-mail:hujun@cqupt.edu.cn
  • 作者简介:E⁃mail:hujun@cqupt.edu.cn
  • 基金资助:
    国家自然科学基金(61772096);重庆市自然科学基金(cstc2019jcyj?cxttX0002);重庆市教委重点合作项目(HZ2021008)

A dynamic attribute reduction based on three⁃way decisions for multi⁃user preferences

Xin Liu, Jun Hu(), Qinghua Zhang, Hong Yu   

  1. Chongqing Key Laboratory of Computational Intelligence, Chongqing University of Posts and Telecommunications, Chongqing, 400065, China
  • Received:2021-06-23 Online:2022-01-30 Published:2022-02-22
  • Contact: Jun Hu E-mail:hujun@cqupt.edu.cn

摘要:

属性约简是粗糙集理论的核心研究内容之一.为满足不同用户对约简的不同需求,针对多用户偏好改变的情形,提出一种面向多用户的三支动态属性约简方法.首先,融合多用户偏好,定义用户偏好矩阵描述多用户下各属性偏好度;然后,结合属性偏好度和现实问题的代价,提出用户偏好指标,表示属性在当前用户组下的重要程度,并作为启发信息选择属性;最后,利用三支决策理论对约简集合和非约简集合进行属性的三分而治,达到更新约简的目的.实例分析及实验结果验证了方法的可行性和有效性,并且得到的约简能较好地满足多用户需求.

关键词: 属性约简, 用户偏好, 三支决策, 约简更新

Abstract:

Attribute reduction is one of the key issues of rough set theory. In order to satisfy different users' requirements for reduction,a dynamic attribute reduction method based on three?way decisions is proposed to adapt to the change of user preferences. First,incorporating multi?user preferences,a user preference matrix is defined to describe the preference of each attribute under multiple users. Then,combining attribute preference degree and cost in real problems,the user preference index is proposed to indicate the importance of attribute in the current user group,and used as heuristic information for attribute selection. Finally,the three?way decisions theory is used to divide the reduced set and non?reduced set into three parts so as to achieve the purpose of updating reduction by taking different strategy. Case analysis and experimental results verify the feasibility and effectiveness of the proposed method,and the obtained reduction can satisfy the requirement of multi?user well.

Key words: attribute reduction, users' preference, three?way decisions, reduction updating

中图分类号: 

  • TP18

图1

三支决策模型"

表1

决策表"

a1a2a3a4a5D
211111
111211
120121
121112
012112
211322
221122

表2

属性代价矩阵"

属性a1a2a3a4a5
ta61281510

图2

属性三分过程"

图3

多用户下动态约简算法流程图"

表3

实验中使用的UCI数据集"

数据集样本数量属性数量
Lymphography14818
Lung?cancer3256
Dermatology36634
Connect?449742

表4

Lymphography数据集上的约简更新过程"

UserIDReductUserIDReduct
1a5,a14,a12,a2,a6,a11,a8,a18,a110a12,a15,a5,a13,a2,a10,a14
2a3,a15,a5,a11,a19a17,a18,a5,a10
3---8a16,a12,a17
4a11,a10,a2,a67a5,a1,a11,a18,a16
5a2,a6,a11,a106a18,a12,a17,a5,a1,a11
6a17,a13,a2,a85a1,a12
7a1,a3,a174a12,a1
8---3---
9---2---
10a3,a11,a1,a61a3,a11,a17,a2
Finala3,a12,a15,a18,a13,a14,a11Finala3,a12,a15,a18,a13,a14,a11

表5

Lung?cancer数据集上的约简更新过程"

UserIDReductUserIDReduct
1b17,b21,b30,b12,b33,b28,b53,b2710b18,b33,b32,b43,b5,b10,b56,b28,b16
2b38,b5,b55,b52,b8,b17,b21,b30,b53,b279b12,b38,b3,b18,b32,b43,b56,b28,b16
3b40,b15,b6,b33,b38,b28,b55,b88b18,b21,b7,b11,b16,b38,b10,b3
4b46,b30,b16,b40,b157b38,b55,b8,b10,b21,b7,b11,b16
5b21,b38,b12,b166b7,b2,b38,b8,b10
6---5b38,b32,b30,b25,b18,b5,b12,b55,b2
7b33,b12,b5,b46,b214---
8b5,b21,b46,b33,b123b18,b5,b21,b12,b10,b32,b7,b25
9---2b52,b15,b6,b18,b38,b30,b12,b10
10b33,b32,b5,b461b38,b32,b30,b5,b15
Finalb33,b21,b38,b32,b52,b30,b6Finalb33,b21,b38,b32,b52,b30,b6

表6

Dermatology数据集上的约简更新过程"

UserIDReductUserIDReduct
1c33,c17,c16,c19,c21,c9,c4,c5,c1810c14,c4,c1,c32,c26,c18,c34
2c8,c1,c34,c33,c16,c9,c5,c189c12,c3,c18
3c3,c33,c26,c8,c48---
4c29,c16,c33,c17c33,c10,c5,c9,c28,c12,c34
5c4,c16,c266c24,c21,c18,c10,c26,c5,c28
6---5---
7c8,c9,c17,c294c8,c33
8c17,c29,c1,c18,c23,c32,c8,c343c12,c8
9c33,c15,c34,c3,c29,c19,c9,c18,c23,c322c17,c34,c12,c1,c9,c24,c18,c14
10c3,c32,c33,c1,c151---
Finalc3,c17,c4,c21,c32,c34Finalc3,c17,c4,c21,c32,c34

表7

Connect?4数据集上的约简更新过程"

UserIDReductUserIDRecuct
1d8,d16,d26,d24,d2,d7,d3,d25,d37,d13,d39,d15,d27,d14,d22,d38,d1,d17,d3110d15,d27,d1,d7,d38,d37,d33,d8,d26,d31,d25,d2,d13,d14,d21
2d21,d16,d3,d13,d39,d27,d179d23,d17,d27,d33,d25
3---8d16,d32,d22,d2,d13
4d16,d39,d13,d17,d3,d27,d217d13,d22
5d32,d21,d24,d14,d26,d39,d13,d17,d3,d276---
6d26,d24,d14,d13,d32,d8,d75---
7d23,d7,d24,d224---
8d32,d22,d38,d73---
9d38,d8,d7,d17,d22,d25,d22---
10d25,d171d25,d17
Finald16,d15,d26,d31,d32,d38,d23,d1,d8,d7,d14,d13,d25,d37,d21Finald16,d15,d26,d31,d32,d38,d23,d1,d8,d7,d14,d13,d25,d37,d21

图4

Lymphography数据集上使用不同分类器的分类精度"

图5

Lung?cancer数据集上使用不同分类器的分类精度"

图6

Dermatology数据集上使用不同分类器的分类精度"

图7

Connect?4数据集上使用不同分类器的分类精度"

图8

Lymphography数据集上的平均SAP"

图9

Lung?cancer数据集上的平均SAP"

图10

Dermatology数据集上的平均SAP"

图11

Connect?4数据集上的平均SAP"

1 Pawlak Z. Rough sets. International Journal of Computer and Information Sciences,1982,11(5):341-356.
2 Jia X Y,Shang L,Zhou B,et al. Generalized attribute reduct in rough set theory. Knowledge?Based Systems,2016(91):204-218.
3 Min F,He H P,Qian Y H,et al. Test?cost?sensitive attribute reduction. Information Sciences,2011,181(22):4928-4942.
4 Wang J,Wang J. Reduction algorithms based on discernibility matrix:The ordered attributes method. Journal of Computer Science and Technology,2001,16(6):489-504.
5 Zhao K,Wang J. A reduction algorithm meeting users' requirements. Journal of Computer Science and Technology,2002,17(5):578.
6 韩素青,阴桂梅. 一种面向用户需求的属性约简算法. 模式识别与人工智能,2014,27(3):281-288.
Han S Q,Yin G M. An user?oriented attribute reduct construction algorithm. Pattern Recognition and Artificial Intelligence,2014,27(3):281-288.
7 Yao Y Y,Zhao Y,Wang J,et al. A model of machine learning based on user preference of attributes∥International Conference on Rough Sets and Current Trends in Computing. Springer Berlin Heidelberg,2006:587-596.
8 Yao Y Y,Zhao Y,Wang J,et al. A model of user?oriented reduct construction for machine learning∥Transactions on rough sets VIII. Springer Berlin Heidelberg,2008:332-351.
9 Liang H L,Wang J,Yao Y Y. User?oriented feature selection for machine learning. The Computer Journal,2007,50(4):421-434.
10 官礼和,王国胤,于洪. 属性序下的增量式Pawlak约简算法. 西南交通大学学报,2011,46(3):461-468.
Guan L H,Wang G Y,Yu H. Incremental algorithm of Pawlak reduction based on attribute order. Journal of Southwest Jiaotong University,2011,46(3):461-468.
11 官礼和,王国胤,胡峰. 一种基于属性序的决策规则挖掘算法. 控制与决策,2012,27(2):313-316.
Guan L H,Wang G Y,Hu F. A decision rules mining algorithm based on attribute order. Control and Decision,2012,27(2):313-316.
12 Yue X D,Chen Y F,Qian J,et al. Attributes reduction model with user preferences∥2014 IEEE 7th Joint International Information Technology and Artificial Intelligence Conference. Chongqing,China:IEEE,2014:191-196.
13 Jing Y G,Li T R,Fujita H,et al. An incremental attribute reduction method for dynamic data mining. Information Sciences,2018(465):202-218.
14 Ma F M,Ding M W,Zhang T F,et al. Compressed binary discernibility matrix based incremental attribute reduction algorithm for group dynamic data. Neurocomputing,2019(344):20-27.
15 Wang F,Liang J Y,Dang C Y. Attribute reduction for dynamic data sets. Applied Soft Computing,2013,13(1):676-689.
16 王国胤. Rough集理论与知识获取. 西安:西安交通大学出版社,2001:23-26,133-136.
17 刘盾,李天瑞,杨新,等. 三支决策:基于粗糙集与粒计算研究视角. 智能系统学报,2019,14(6):1111-1120.
Liu D,Li T R,Yang X,et al. Three?way decisions:Research perspectives for rough sets and granular computing. CAAI Transactions on Intelligent Systems,2019,14(6):1111-1120.
18 Yao Y Y. Three?way decision and granular computing. International Journal of Approximate Reasoning,2018(103):107-123.
19 刘鑫,胡军,张清华. 属性组序下基于代价敏感的约简方法. 南京大学学报(自然科学),2020,56(4):469-479.
Liu X,Hu J,Zhang Q H. Attribute reduction based on cost sensitive under attribute group order. Nanjing University (Natural Science),2020,56(4):469-479.
[1] 李同军, 张晓雨, 吴伟志, 谭安辉. 形式模糊背景中单边模糊概念格的属性约简[J]. 南京大学学报(自然科学版), 2022, 58(1): 38-48.
[2] 刘小伟, 景运革. 一种有效更新多源数据约简的增量算法[J]. 南京大学学报(自然科学版), 2021, 57(6): 1083-1091.
[3] 张呈玲, 李进金, 林艺东. 基于面向对象(属性)概念格的三支规则提取[J]. 南京大学学报(自然科学版), 2021, 57(4): 599-610.
[4] 王颖俐, 魏玲. 基于改进的区间损失函数聚合法的三支决策[J]. 南京大学学报(自然科学版), 2021, 57(3): 493-501.
[5] 袁晓峰, 钱苏斌, 周彩根. 基于填充先验约束的矩阵分解算法[J]. 南京大学学报(自然科学版), 2021, 57(2): 197-207.
[6] 杜祥通, 李永忠. 基于深度信念网络和三支决策的入侵检测算法[J]. 南京大学学报(自然科学版), 2021, 57(2): 272-278.
[7] 刘鑫,胡军,张清华. 属性组序下基于代价敏感的约简方法[J]. 南京大学学报(自然科学版), 2020, 56(4): 469-479.
[8] 顾萍萍,周献中. 基于概率语言术语集评价的三支决策方法研究[J]. 南京大学学报(自然科学版), 2020, 56(4): 505-514.
[9] 徐媛媛,张恒汝,闵帆,黄雨婷. 三支交互推荐[J]. 南京大学学报(自然科学版), 2019, 55(6): 973-983.
[10] 程永林, 李德玉, 王素格. 基于极大相容块的邻域粗糙集模型[J]. 南京大学学报(自然科学版), 2019, 55(4): 529-536.
[11] 龙柄翰, 徐伟华. 模糊三支概念分析与模糊三支概念格[J]. 南京大学学报(自然科学版), 2019, 55(4): 537-545.
[12] 张龙波, 李智远, 杨习贝, 王怡博. 决策代价约简求解中的交叉验证策略[J]. 南京大学学报(自然科学版), 2019, 55(4): 601-608.
[13] 张 婷1,2,张红云1,2*,王 真3. 基于三支决策粗糙集的迭代量化的图像检索算法[J]. 南京大学学报(自然科学版), 2018, 54(4): 714-.
[14] 靳义林1,2*,胡 峰1,2. 基于三支决策的中文文本分类算法研究[J]. 南京大学学报(自然科学版), 2018, 54(4): 794-.
[15] 陶玉枝1,2,赵仕梅1,2,谭安辉1,2*. 一种基于决策表约简的集覆盖问题的近似解法[J]. 南京大学学报(自然科学版), 2018, 54(4): 821-.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!