南京大学学报(自然科学版) ›› 2019, Vol. 55 ›› Issue (4): 609–623.doi: 10.13232/j.cnki.jnju.2019.04.011

所属专题: 测试专题

• • 上一篇    下一篇

复杂网络下的概念认知学习

马娜1,2,范敏1,2(),李金海1,2   

  1. 1. 昆明理工大学理学院,昆明,650500
    2. 昆明理工大学数据科学研究中心,昆明,650500
  • 收稿日期:2019-05-28 出版日期:2019-07-30 发布日期:2019-07-23
  • 通讯作者: 范敏 E-mail:fmkmust@163.com
  • 基金资助:
    国家自然科学基金(61562050,61573173,61573321,41631179)

Concept⁃cognitive learning under complex network

Na Ma1,2,Min Fan1,2(),Jinhai Li1,2   

  1. 1. Faculty of Science,Kunming University of Science and Technology,Kunming,650500,China
    2. Data Science Research Center,Kunming University of Science and Technology,Kunming,650500,China
  • Received:2019-05-28 Online:2019-07-30 Published:2019-07-23
  • Contact: Min Fan E-mail:fmkmust@163.com

摘要:

数据分析中,从网络中进行概念认知学习是网络背景下的机器学习或人工智能领域的重要问题.首先通过分析复杂网络方法与形式概念方法的数据基础,将二者的数据通过邻接矩阵与关联矩阵统一起来,提出一种网络形式背景框架,使以上两种理论与方法之间有了互通的桥梁,从而可以结合它们各自的优势对网络概念进行更深入的研究.在此基础上,从网络概念的三个层次出发研究了以下内容:(1)通过定义节点的结构影响力和内涵影响力并将它们进行加权,定义了节点的网络影响力.(2)通过分析扩散网络、收缩网络的特点提出强概念、弱概念、网络概念,并给出了网络概念的特征值:概念的势、概念平均度.于是,该理论不仅能在网络中找到网络概念,还能给出网络概念的重要性和网络概念内部的差异性.(3)研究了强(弱)概念的有关性质,为以后构造相应的代数系统,生成各种网络概念算子提供了理论基础.

关键词: 复杂网络, 形式概念, 网络形式背景, 网络影响力, 网络概念

Abstract:

In data analysis,concept?cognitive learning from the network is an important issue in the field of machine learning or artificial intelligence in the context of network. Firstly,by analyzing the data foundation of complex network and formal concept methods,the data induced by them are unified through the adjacency matrix and the association matrix,and a formal context of network framework is proposed. It is a bridge between the above two theories and methods,so that the research of network concept will be more deeply and meaningful by combining their respective advantages. On this basis,the following contents are studied from three hierarchies of the network concept:(1)by defining the structural influence and connotation influence of the nodes and weighting them,the network influence of the nodes is defined. (2)By analyzing the characteristics of diffusion network and convergent network,strong concept,weak concept,network concept are defined,meanwhile the eigenvalues ??of network concept: concept average diversity,concept average degree are given. Therefore,not only can the network concept be found in the network by the theory,but also the importance of the network concept and the internal differences of the network concept will be given by the theory.

(3) The related properties of the strong (weak) concept are studied,which provides a theoretical basis for constructing the corresponding algebraic systems and generating various network concept operators.

Key words: complex network, formal concept, network formal context, network influence, network concept

中图分类号: 

  • TP18

表1

网络形式背景(U,M,A,I)"

M1MkA
x1x2xnx1x2xna1a2am
x1101100110
x2010001010
xn101011011

图1

网络图"

表2

图1的网络形式背景"

M1M2A
123456123456abcde
100101101010010100
200011110010101000
310000000001111010
401000111001110000
511000000110100111
611010001111010111

表3

不同α下六位作者的网络影响力"

α0.10.20.30.40.50.60.70.80.9
ω10.57740.59930.62110.64290.66470.68660.70840.73020.7521
ω20.08700.17390.26090.34780.43480.52180.60870.69570.7826
ω30.44080.43730.43370.43010.42650.42300.41940.41580.4123
ω40.38430.43530.48640.53740.58840.63940.69040.74150.7925
ω50.56960.58360.59760.61160.62570.63970.65370.66770.6817
ω61.00001.00001.00001.00001.00001.00001.00001.00001.0000

图2

科研网络"

表4

科研网络的网络形式背景"

M1A
1234abcd
101011101
210110100
301010011
411101110

表5

共同属性矩阵"

1234
1bdab
2bb
3dc
4abbc
1 MoroneF,MakseH. Influence maximization in complex networks through optimal percolation. Nature,2015,524(7579):65-68.
2 WengL L,MenczerF,AhnY Y. Virality prediction and community structure in social networks. Scientific Reports,2013,3:2522.
3 GirvanM,NewmanM E J. Community structure in social and biological networks. Proceedings of the National Academy of Sciences of the United States of America,2002,99(12):7821-7826.
4 GuimeràR,Sales?PardoM. Missing and spurious interactions and the reconstruction of complex networks. Proceedings of the National Academy of Sciences of the United States of America,2010,106(52):22073-22078.
5 KumarC A,IshwaryaM S,LooC K. Formal concept analysis approach to cognitive functionalities of bidirectional associative memory. Biologically Inspired Cognitive Architectures,2015,12:20-33.
6 XuW H,LiW T. Granular computing approach to two?way learning based on formal concept analysis in fuzzy datasets. IEEE Transactions on Cybernetics,2016,46(2):366-379.
7 LiJ H,HuangC C,QiJ J,et al. Three?way cognitive concept learning via multi?granularity. Information Sciences,2017,378:244-263.
8 SunZ J,WangB,ShengJ F,et al. Identifying influential nodes in complex networks based on weighted formal concept analysis. IEEE Access,2017,5:3777-3789.
9 PengS C,YangA M,CaoL H,et al. Social influence modeling using information theory in mobile social networks. Information Sciences,2017,379:146-159.
10 JaliliM,PercM. Information cascades in complex networks. Journal of Complex Networks,2017,5(5):665-693.
11 徐伟华,李金海,魏玲等. 形式概念分析理论与应用. 北京:科学出版社,2016,
Xu W H,Li J H,Wei L,et al. Formal concept analysis:theory and application. Beijing:Science Press,2016,263.
12 NewmanM E J.Networks:An Introduction. Oxford:Oxford University Press,2010,784.
13 林聚任. 社会网络分析:理论、方法与应用. 北京:北京师范大学出版社,2009,306.
Lin J R.Social network analysis:theory, method and application. Beijing: Beijing Normal University Press,2009,306.
14 廖丽平,胡仁杰,张光宇. 模糊社会网络的中心度分析方法. 模糊系统与数学,2013,27(2):169-176.
Liao L P,Hu R J,Zhang G Y.The centrality analysis of fuzzy social network. Fuzzy Systems and Mathematics,2013,27(2):169-176.
15 SabidussiG. The centrality index of a graph. Psychometrika,1996,31(4):581-603.
16 BonacichP. Factoring and weighting approaches to status scores and clique identification. The Journal of Mathematical Sociology,1972,2(1):113-120.
17 GanterB,WilleR. Formal concept analysis:mathematical foundations. Springer Berlin Heidelberg,1999,287.
[1] 李俊余, 李星璇, 王霞, 吴伟志. 基于三元因子分析的三元概念约简[J]. 南京大学学报(自然科学版), 2020, 56(4): 480-493.
[2] 刘胜久,李天瑞,珠杰,刘佳. 带权图的多重分形研究[J]. 南京大学学报(自然科学版), 2020, 56(1): 85-97.
[3] 郑文萍,刘韶倩,穆俊芳. 一种基于相对熵的随机游走相似性度量模型[J]. 南京大学学报(自然科学版), 2019, 55(6): 984-999.
[4] 钱 峰1,2,张 蕾1,2,赵 姝1*,陈 洁1,张燕平1. 基于加权树的层次社团划分算法[J]. 南京大学学报(自然科学版), 2018, 54(4): 696-.
[5] 李俊余1,2,朱荣杰1,王 霞1,2*,吴伟志1,2. 三元概念与形式概念的关系[J]. 南京大学学报(自然科学版), 2018, 54(4): 786-.
[6]  张泽华1*,段力畑1,段 富1,张 楠2.  基于局部结构特征的重叠社区挖掘研究进展[J]. 南京大学学报(自然科学版), 2017, 53(3): 537-.
[7] 宾 晟*,孙更新. 基于多子网复合复杂网络模型的多关系社交网络重要节点发现算法[J]. 南京大学学报(自然科学版), 2017, 53(2): 378-.
[8] 张燕平1,2,汪  洋1,2,赵  姝1,2, 段  震1,2**. 基于覆盖的社团发现算法[J]. 南京大学学报(自然科学版), 2013, 49(5): 539-545.
[9] 郭铭铭,窦建华,杨彬
. 基于形式化概念分析和概念相似性度量的程序重组方法
[J]. 南京大学学报(自然科学版), 2011, 47(5): 594-604.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
[1] 秦 娅, 申国伟, 赵文波, 陈艳平. 基于深度神经网络的网络安全实体识别方法[J]. 南京大学学报(自然科学版), 2019, 55(1): 29 -40 .
[2] 郭小松,赵红丽,贾俊芳,杨静,孟祥军. 密度泛函理论方法研究第一系列过渡金属对甘氨酸的配位能力[J]. 南京大学学报(自然科学版), 2019, 55(6): 1040 -1046 .
[3] 党政,代群威,安超,彭启轩,卓曼他,杨丽君. 静态水蚀条件下自然钙华预制块的溶出特性研究[J]. 南京大学学报(自然科学版), 2019, 55(6): 916 -923 .
[4] 韩普,刘亦卓,李晓艳. 基于深度学习和多特征融合的中文电子病历实体识别研究[J]. 南京大学学报(自然科学版), 2019, 55(6): 942 -951 .
[5] 郑文萍,刘韶倩,穆俊芳. 一种基于相对熵的随机游走相似性度量模型[J]. 南京大学学报(自然科学版), 2019, 55(6): 984 -999 .
[6] 吕国俊,曹建军,郑奇斌,常宸,翁年凤. 基于结构保持对抗网络的跨模态实体分辨[J]. 南京大学学报(自然科学版), 2020, 56(2): 197 -205 .
[7] 汪洋,陈泰格,陆晓凡,辛小燕,王坤,李茗,青钊,张英为,严晓敏,吴超,言方荣,张冰. COVID⁃19的临床和影像特征与试行指南的映证分析[J]. 南京大学学报(自然科学版), 2020, 56(3): 430 -436 .
[8] 郑建兴,李沁文,王素格,李德玉. 基于翻译模型的异质重边信息网络链路预测研究[J]. 南京大学学报(自然科学版), 2020, 56(4): 541 -548 .