南京大学学报(自然科学版) ›› 2022, Vol. 58 ›› Issue (1): 82–93.doi: 10.13232/j.cnki.jnju.2022.01.009

• • 上一篇    下一篇

基于极大相容块的不完备信息处理新方法及其应用

王敬前1, 张小红1,2()   

  1. 1.陕西科技大学数学与数据科学学院, 西安, 710021
    2.陕西省人工智能联合实验室, 陕西科技大学, 西安, 710021
  • 收稿日期:2021-07-12 出版日期:2022-01-30 发布日期:2022-02-22
  • 通讯作者: 张小红 E-mail:zhangxiaohong@sust.edu.cn
  • 作者简介:E⁃mail:zhangxiaohong@sust.edu.cn
  • 基金资助:
    国家自然科学基金(61976130)

A new method of incomplete information processing based on maximal consistent block and its application

Jingqian Wang1, Xiaohong Zhang1,2()   

  1. 1.School of Mathematics and Data Science, Shanxi University of Science and Technology, Xi'an, 710021, China
    2.Shanxi Joint Laboratory of Artificial Intelligence, Shanxi University of Science and Technology, Xi'an, 710021, China
  • Received:2021-07-12 Online:2022-01-30 Published:2022-02-22
  • Contact: Xiaohong Zhang E-mail:zhangxiaohong@sust.edu.cn

摘要:

针对不完备信息提出一种新的基于矩阵方法的极大相容块求取算法与属性约简方法,结合智能分类器给出不完备信息条件下的故障诊断方法.首先,通过矩阵方法计算不完备决策表中的极大相容块;然后,利用所求得的极大相容块,提出一种新的属性约简算法,并与其他方法做对比;最后,将所提出的基于极大相容块的属性约简方法与智能分类器(支持向量机、随机森林、决策树等)结合,建立优化的智能故障分类器,将它应用于不完备信息条件下的故障诊断.以汽轮机组的故障诊断为例进行仿真实验,实验结果表明提出的针对不完备信息条件下的故障诊断方法可行、有效.

关键词: 极大相容块, 覆盖粗糙集, 矩阵方法, 不完备信息, 故障诊断

Abstract:

In this paper,a new maximum compatible block algorithm and attribute reduction method based on matrix method are proposed for incomplete information,and the fault diagnosis method under the condition of incomplete information is given by the maximal consistent block and intelligent classifiers. Firstly,the maximal consistent block in an incomplete decision table is calculated by matrices. Then,a new attribute reduction algorithm is proposed based on the maximal consistent block and compared with other methods. Finally,some optimized intelligent fault classifiers are established by the combination between the proposed attribute reduction method and corresponding classifiers,such as support vector machine,random forest and decision tree. It is applied to fault diagnosis under the condition of incomplete information. Moreover,the fault diagnosis of a steam turbine as an example for the simulation. Experimental results shows that the proposed method is feasible and effective.

Key words: maximal consistent block, covering?based rough set, matrix approach, incomplete information, fault diagnosis of steam turbine

中图分类号: 

  • TP30

表1

不完备决策表DT[14-15]"

UabcdefghD
x1*00**0201
x2*2**1*010
x3*2*1*2011
x4*2*112011
x51***10*00
x62320*1310
x723201*311
x8232131*11
x93**3102*0
x10321**0230
x11321110**0
x12321311211

表2

任意对象xi∈U的所有极大相容块CxiMi=1,2,?,12"

CxiMi=1,2,?,6CxiMi=7,8,?,12
Cx1Mx1,x5,x1,x9Cx7Mx6,x7
Cx2Mx2,x11,x2,x3,x4Cx8Mx8
Cx3Mx2,x3,x4Cx9Mx1,x9,x9,x10
Cx4Mx2,x3,x4Cx10Mx9,x10,x10,x11
Cx5Mx1,x5Cx11Mx2,x11,x10,x11
Cx6Mx6,x7Cx12Mx12

表3

DT的极大相容块最全描述决策表 (CDT)"

abcdefghD'
X1100*10200, 1
X2300310200, 1
X3*2*112010, 1
X4321110010
X5232011310, 1
X6232131*11
X7321310230
X8321110230
X9321311211

表4

本文算法与其他不完备决策表属性约简算法的比较"

约简算法1[15]约简算法2[14]本文算法
例1的约简a,c,d,e,f,ha,d,f,hd,f,h
基于容差关系
基于极大相容块
基于分辨矩阵
转换为新的决策表
考虑决策属性

图1

“极大相容块+智能分类器”的故障诊断方法流程"

表5

不完备汽轮机故障信息表"

UabcdeD
10.0520.7830.225*0.0131
20.2320.9750.3140.056*1
30.161*0.2850.0230.0161
40.1060.858*0.0170.0281
5*0.8190.2010.0160.0121
60.0280.0610.98*0.0572
70.0450.022*0.3160.0652
80.010.0540.8750.183*2
9*0.0320.9230.2190.0372
100.023*0.7580.1150.0192
110.0330.0370.3860.5310.233
12*0.023*0.4580.1033
130.012*0.4270.4960.1753
140.0210.0170.2980.403*3
150.0170.0560.483*0.3013

表6

离散化断点表"

M012
a0,0.01970.0197,0.04370.0437,1
b0,0.03530.0353,0.3020.302,1
c0,0.3090.309,0.5750.575,1
d0,0.09530.0953,0.3450.345,1
e0,0.0250.025,0.0830.083,1

表7

不完备汽轮机故障诊断决策表"

UabcdeD
1110*01
21100*1
31*0001
411*001
5*10001
6001*02
700*002
80010*2
9*01002
100*1002
11000113
12*0*113
130*0113
140001*3
15000*13

表8

关于条件属性集M的极大相容块C(M)"

X11,3,5X610
X22,4X711
X36X812,13
X47,9X912,14
X58X1013,15

表9

完备化的数据集"

UabcdeD
10.0520.7830.2250.0280.0131
20.2320.9750.3140.0560.0171
30.1610.8590.2850.0230.0161
40.1060.8580.2560.0170.0281
50.1380.8190.2010.0160.0121
60.0280.0610.980.2080.0572
70.0450.0220.8840.3160.0652
80.0100.0540.8750.1830.0452
90.0270.0320.9230.2190.0372
100.0230.0420.7580.1150.0192
110.0330.0370.3860.5310.233
120.0210.0230.3990.4580.1033
130.0120.0330.4270.4960.1753
140.0210.0170.2980.4030.2023
150.0170.0560.4830.4720.3013

表10

汽轮机组故障的测试集[17-18]"

UabcdeD
160.1610.7530.1280.0060.0031
170.2820.9050.3430.0460.0281
180.0170.0530.750.2520.1072
190.0450.2220.9890.3860.0152
200.0270.1270.8510.6190.2522
210.0260.0430.3570.5170.0983
220.0230.1370.3780.4210.1523
230.01840.00550.82050.1270.00313

图2

“极大相容块+SVM”和SVM对测试集的故障分类图"

表11

“极大相容块+SVM”和SVM对测试集的故障分类情况"

极大相容块+SVMSVM
准确率运行时间(s)准确率运行时间(s)
Polynomial87.5%10.6375%11.15
RBF87.5%14.3375%15.56
Sigmoid87.5%12.6475%12.89

表12

其他类型的“极大相容块+智能分类器”的故障分类情况"

极大相容块+智能分类器准确率
极大相容块+随机森林87.5%
极大相容块+决策树87.5%
随机森林75%
决策树62.5%
1 陈泽华,宋波,闫继雄,等. 基于概念格的不完备信息系统最简规则提取算法. 控制与决策,2019,34(5):1011-1017.
Chen Z H,Song B,Yan J X,et al. Concise rule extraction algorithm of incomplete information system based on concept lattice. Control and Decision,2019,34(5):1011-1017.
2 Pawlak Z. Rough sets. International Journal of Computer and Information Science,1982,11(5):341-356.
3 Chen D G,Dong L J,Mi J S. Incremental mechanism of attribute reduction based on discernible relations for dynamically increasing attribute. Soft Computing,2020,24(1):321-332.
4 Tawhid M A,Ibrahim A M. Feature selection based on rough set approach,wrapper approach,and binary whale optimization algorithm. International Journal of Machine Learning and Cybernetics,2020,11(3):573-602.
5 韩朝,苗夺谦,任福继,等. 基于粗糙集知识发现的开放领域中文问答检索. 计算机研究与发展,2018,55(5):958-967.
Han Z,Miao D Q,Ren F J,et al. Rough set knowledge discovery based open domain Chinese question answering retrieval. Journal of Computer research and Development,2018,55(5):958-967.
6 姚晟,汪杰,徐风,等. 不完备邻域粗糙集的不确定性度量和属性约简. 计算机应用,2018,38(1):97-103.
Yao S,Wang J,Xu F,et al. Uncertainty measurement and attribute reduction in incomplete neighborhood rough set. Journal of Computer Applications,2018,38(1):97-103.
7 Zheng J,Wang Y M,Zhang K,et al. A dynamic emergency decision?making method based on group decision making with uncertainty information. International Journal of Disaster Risk Science,2020,11(5):667-679.
8 Yang L,Zhang X Y,Xu W H,et al. Multi?granulation rough sets and uncertainty measurement for multi?source fuzzy information system. International Journal of Fuzzy Systems,2019,21(6):1919-1937.
9 孙妍,米据生,冯涛,等. 变精度极大相容块粗糙集模型及其属性约简. 计算机科学与探索,2020,14(5):892-900.
Sun Y,Mi J S,Feng T,et al. Maximum consistent block based variable precision rough set model and attribute reduction. Journal of Frontiers of Computer Science & Technology,2020,14(5):892-900.
10 Sang B B,Yang L,Chen H M,et al. Generalized multi?granulation double?quantitative decision?theoretic rough set of multi?source information system. International Journal of Approximate Reasoning,2019,115:157-179.
11 杨霁琳,秦克云,裴峥. 不完备决策表中基于相似关系的属性约简. 计算机工程,2010,36(20):10-12.
Yang J L,Qin K Y,Pei Z. Attributes reduction based on similarity relation in incomplete decision table. Computer Engineering,2010,36(20):10-12.
12 Kryszkiewicz M. Rough set approach to incomplete information systems. Information Sciences,1998,112(1-4):39-49.
13 Leung Y,Li D Y. Maximal consistent block technique for rule acquisition in incomplete information systems. Information Sciences,2003(153):85-106.
14 黄治国,王淼. 不完备决策表中基于容差关系的属性约简方法. 微电子学与计算机,2016,33(6):147-151,156.
Huang Z G,Wang M. Attribute reduction based on tolerance relation in incomplete decision table. Microelectronics & Computer,2016,33(6):147-151,156.
15 鄂旭,邵良杉,周津,等. 一种新的不完备信息系统属性约简算法. 重庆邮电大学学报(自然科学版),2010,22(5):648-651,664.
E X,Shao L S,Zhou J,et al. A new attribute reduction algorithm in an incomplete information system. Journal of Chongqing University of Posts and Telecommunications (Natural Science Edition),2010,22(5):648-651,664.
16 黄文涛,赵学增,王伟杰,等. 基于不完备数据的汽轮机组故障诊断的粗糙集方法. 汽轮机技术,2004,46(1):57-59.
Huang W T,Zhao X Z,Wang W J,et al. Rough sets method of fault diagnosis for steam turbine based on incomplete data. Turbine Technology,2004,46(1):57-59.
17 李化. 汽轮发电机组振动故障智能诊断模型的理论及方法研究. 博士学位论文. 重庆:重庆大学,1999.
Li H. Research on theory and method of intelligent diagnosis model for turbine generator vibration fault. Ph.D. Dissertation. Chongqing:Chongqing University,1999.
18 邴汉昆. 基于小波分析和SVM的汽轮机非线性振动故障诊断研究. 硕士学位论文. 北京:华北电力大学,2013.
Bing H K. Research on nonlinear vibration fault diagnosis of steam turbine based on wavelet analysis and SVM. Master Dissertation. Beijing:North China Electric Power University,2013.
[1] 陈石,张兴敢. 基于小波包能量熵和随机森林的级联H桥多电平逆变器故障诊断[J]. 南京大学学报(自然科学版), 2020, 56(2): 284-289.
[2] 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.
[3] 安 晶, 艾 萍, 徐 森, 刘 聪, 夏建生, 刘大琨. 一种基于一维卷积神经网络的旋转机械智能故障诊断方法[J]. 南京大学学报(自然科学版), 2019, 55(1): 133-142.
[4]  李 敬,王利东*.  不完备信息系统中的广义多粒度双相对定量决策粗糙集[J]. 南京大学学报(自然科学版), 2017, 53(4): 782-.
[5] 杨春梅1*,骆文于2,乔方利1,吕连港1. 均匀可穿透楔形波导中的三维声场解析解[J]. 南京大学学报(自然科学版), 2015, 51(6): 1210-1216.
[6] 马兴斌1,鞠恒荣1,2,3,杨习贝1,2,3,宋晶晶1,2,3. 不完备信息系统中的多重代价决策粗糙集[J]. 南京大学学报(自然科学版), 2015, 51(2): 335-342.
[7] 汤建国1,汪江桦1,佘堃2,祝峰3. 不同覆盖产生相同覆盖近似集的条件研究[J]. 南京大学学报(自然科学版), 2014, 50(1): 72-.
[8] 顾沈明,吴伟志,徐优红. 不完备多标记信息系统中粒度研究[J]. 南京大学学报(自然科学版), 2013, 49(5): 567-573.
[9]  顾沈明**,叶晓敏,吴伟志
.  多标记粒度不完备信息系统的粗糙近似*[J]. 南京大学学报(自然科学版), 2013, 49(2): 250-257.
[10]  赵荣泳 1 , 李翠玲 2 ** , 高晓康 3 , 王昭云 4
.  电子镇流器故障诊断的变精度粗糙集模型*
[J]. 南京大学学报(自然科学版), 2010, 46(5): 494-500.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!