Please wait a minute...
浙江大学学报(工学版)  2020, Vol. 54 Issue (11): 2247-2257    DOI: 10.3785/j.issn.1008-973X.2020.11.020
生物医学工程     
基于鱼群算法的脑功能连接邻域粗糙集特征归约方法
冀俊忠1,2(),宋晓妮1,2,杨翠翠1,2,*()
1. 北京工业大学 信息学部,多媒体与智能软件技术北京市重点实验室,北京 100124
2. 北京工业大学 北京人工智能研究院,北京 100124
Feature reduction of neighborhood rough set based on fish swarm algorithm in brain functional connectivity
Jun-zhong JI1,2(),Xiao-ni SONG1,2,Cui-cui YANG1,2,*()
1. Beijing Municipal Key Laboratory of Multimedia and Intelligent Software Technology, Faculty of Information Technology, Beijing University of Technology, Beijing 100124, China
2. Beijing Institute of Artificial Intelligence, Beijing University of Technology, Beijing 100124, China
 全文: PDF(882 KB)   HTML
摘要:

为了有效应对脑功能连接高维小样本性给分类模型构建带来的挑战,得到与脑疾病诊断相关的重要特征,提出基于鱼群算法的脑功能连接邻域粗糙集特征归约方法.该方法建立脑功能连接数据的邻域决策表;依据特征的依赖度将鱼个体初始化为候选的脑功能连接特征子集,并采用综合特征子集依赖度和特征子集长度的适应度函数对鱼个体进行评价;在种群优化过程中,执行觅食、聚集、追尾机制,以及交叉和迁徙2个新机制来不断搜索最优的特征子集.在3种脑疾病功能磁共振脑成像(fMRI)数据集上,将所提方法与多种已有的特征归约方法进行对比实验.结果表明,该方法是有效的脑功能连接特征归约方法,可以有效降低脑功能连接数据的维度,获得分类判别能力较强的脑功能连接特征.

关键词: 脑功能连接特征归约鱼群算法邻域粗糙集    
Abstract:

Feature reduction of neighborhood rough set based on fish swarm algorithm in brain functional connectivity was proposed, in order to effectively deal with the challenge brought by the high-dimensional and small sample size of brain functional connectivity to the construction of classification model, and to obtain important features related to brain disease diagnosis. The neighborhood decision table of the brain functional connectivity is established in the algorithm. Each artificial fish is initialized as a candidate feature subset of the brain functional connectivity according to the feature dependence information, and a fitness function is constructed based on the information of feature subset dependence and feature subset length to evaluate each individual. The preying, swarming, following mechanisms of the fish swarm algorithm, as well as two new simulation mechanisms of crossover and migration are performed to iteratively search for the optimal feature subset in the process of population optimization. The proposed method was compared with a variety of existing feature reduction methods in the functional magnetic resonance imaging (fMRI) data sets of three brain diseases. Results show that the new method is an effective feature reduction method for the brain functional connectivity, which can effectively reduce the dimension of the brain functional connectivity data and obtain the brain functional connectivity features with high classification discrimination ability.

Key words: brain functional connectivity    feature reduction    fish swarm algorithm    neighborhood rough set
收稿日期: 2019-11-11 出版日期: 2020-12-15
CLC:  TP 301  
基金资助: 国家自然科学基金资助项目(61672065,61906010);北京市教委科技资助项目(KM202010005032);中国博士后科学基金资助项目(2018M631291);北京市博士后工作经费资助项目(2017-ZZ-024);朝阳区博士后工作经费资助项目(2018ZZ-01-05)
通讯作者: 杨翠翠     E-mail: jjz01@bjut.edu.cn;yangcc@bjut.edu.cn
作者简介: 冀俊忠(1969—),男,教授,博士,从事生物信息学研究. orcid.org/0000-0001-6951-741X. E-mail: jjz01@bjut.edu.cn
服务  
把本文推荐给朋友
加入引用管理器
E-mail Alert
作者相关文章  
冀俊忠
宋晓妮
杨翠翠

引用本文:

冀俊忠,宋晓妮,杨翠翠. 基于鱼群算法的脑功能连接邻域粗糙集特征归约方法[J]. 浙江大学学报(工学版), 2020, 54(11): 2247-2257.

Jun-zhong JI,Xiao-ni SONG,Cui-cui YANG. Feature reduction of neighborhood rough set based on fish swarm algorithm in brain functional connectivity. Journal of ZheJiang University (Engineering Science), 2020, 54(11): 2247-2257.

链接本文:

http://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2020.11.020        http://www.zjujournals.com/eng/CN/Y2020/V54/I11/2247

数据集 组别 人数 年龄范围
被试
ADNI NC 53 31 22 65~96
MCI 61 29 32 63~88
AD 66 27 39 56~88
ADHD NC 48 27 21 8~16
ADHD 32 28 4 8~15
ABIDE NC 52 39 13 8~56
ASD 48 37 11 8~55
表 1  静息态fMRI数据集
图 1  NRSFSA算法在不同测试参数下的评价指标
图 2  NRSFSA算法在不同邻域半径下的实验结果
图 3  NRSFSA算法在数据集上的迭代过程
数据集 算法 Nu Acc Pr Re Fm
ADHD SRS 23 0.60750 0.58638 0.48067 0.41665
SNRS 11 0.66250 0.64095 0.48600 0.48524
mRMR 9 0.72500 0.75757 0.51510 0.54380
NRS 11 0.69750 0.68933 0.53933 0.52734
LLE 22 0.58250 0.63433 0.40365 0.27369
PCA 39 0.61750 0.68170 0.28600 0.31450
F-score 9 0.64750 0.62633 0.48133 0.48689
NRSFSA 9 0.76500 0.79333 0.61962 0.64165
ABIDE SRS 43 0.53400 0.70633 0.19945 0.24923
SNRS 14 0.51200 0.52567 0.37864 0.40008
mRMR 10 0.56200 0.58071 0.52614 0.50408
NRS 6 0.65800 0.67351 0.68288 0.63672
LLE 36 0.57800 0.56712 0.61083 0.55199
PCA 66 0.53800 0.55890 0.51870 0.49210
F-score 10 0.72600 0.74643 0.68048 0.68616
NRSFSA 10 0.69400 0.73376 0.61683 0.61745
NC-AD SRS 15 0.59152 0.75000 0.07386 0.10269
SNRS 14 0.70924 0.73227 0.74110 0.71883
mRMR 8 0.70167 0.72543 0.78083 0.73496
NRS 17 0.81803 0.80314 0.89052 0.83404
LLE 24 0.71970 0.74294 0.78606 0.74340
PCA 60 0.76920 0.76320 0.82900 0.77960
F-score 8 0.64288 0.66526 0.74277 0.68361
NRSFSA 8 0.75348 0.77167 0.81725 0.77346
NC-MCI SRS 15 0.62379 0.65169 0.71285 0.65813
SNRS 12 0.67530 0.69825 0.76487 0.70573
mRMR 9 0.69712 0.71774 0.76529 0.71338
NRS 15 0.72015 0.72979 0.79019 0.73919
LLE 24 0.73076 0.72760 0.83533 0.76117
PCA 60 0.76710 0.76270 0.84500 0.78490
F-score 9 0.81652 0.82566 0.82013 0.80176
NRSFSA 9 0.75697 0.75238 0.81108 0.76553
AD-MCI SRS 16 0.63077 0.64118 0.71899 0.65242
SNRS 19 0.66179 0.65876 0.77476 0.68663
mRMR 14 0.81859 0.79769 0.87512 0.82196
NRS 16 0.79244 0.80323 0.81253 0.79404
表 2  8种算法在5组实验数据上的实验结果
数据集 算法 Nu Acc Pr Re Fm
LLE 24 0.81936 0.82191 0.81978 0.80961
PCA 64 0.78720 0.78090 0.82660 0.79100
F-score 14 0.81077 0.82612 0.80229 0.80196
NRSFSA 14 0.79231 0.77573 0.86441 0.80501
平均值 SRS 22.4 0.59751 0.66712 0.43716 0.41582
SNRS 14.0 0.64417 0.65118 0.62908 0.59930
mRMR 10.0 0.70088 0.71583 0.69250 0.66364
NRS 13.0 0.73722 0.73980 0.74309 0.70627
LLE 26.0 0.68606 0.69878 0.69113 0.62797
PCA 57.8 0.69580 0.70950 0.66110 0.63240
F-score 10.0 0.72873 0.73796 0.70540 0.69208
NRSFSA 10.0 0.75235 0.76538 0.74584 0.72062
表 2  
数据集 算法
SRS SNRS mRMR NRS LLE PCA F-score NRSFSA
ADHD 1605.45 2428.04 554.124 55.551 3 45.905 0 70.5156 31.9536 2834.76
ABIDE 2083.90 3180.85 571.607 60.0268 75.5599 142.769 0 43.7131 4814.16
NC-AD 2836.60 3190.06 669.555 177.550 0 95.3279 177.166 0 62.1608 4465.16
NC-MCI 3014.95 4081.99 586.957 162.096 0 82.5792 125.213 0 53.4089 5391.35
AD-MCI 4006.34 4785.98 610.880 198.639 0 93.1654 194.548 0 73.9587 11228.40
平均值 2709.45 3533.38 598.625 130.773 0 78.5075 142.042 0 53.0390 5746.77
表 3  8种算法在数据集上运行时间的比较
1 颜志雄, 刘勋, 谭淑平 发展认知神经科学: 人脑毕生发展的功能连接组学时代[J]. 科学通报, 2016, 61 (7): 718- 727
YAN Zhi-xiong, LIU Xun, TAN Shu-ping Developmental cognitive neuroscience: functional connectomics agenda for human brain lifespan development[J]. Chinese Science Bulletin, 2016, 61 (7): 718- 727
doi: 10.1360/N972015-01146
2 WANG S, ZHANG Y, LV L, et al Abnormal regional homogeneity as a potential imaging biomarker for adolescent-onset schizophrenia: aresting-state fMRI study and support vector machine analysis[J]. Schizophrenia Research, 2018, 192: 179- 184
doi: 10.1016/j.schres.2017.05.038
3 COHEN J D, DAW N, WNGELHRSDT B Computational approaches to fMRI analysis[J]. Nature Neuroscience, 2017, 20 (3): 304- 313
doi: 10.1038/nn.4499
4 BASSETT D S, SPORNS O Network neuroscience[J]. Nature Neuroscience, 2017, 20 (3): 353
doi: 10.1038/nn.4502
5 GENG X F, XU J H, LIU B L, et al Multivariate classification of major depressive disorder using the effective connectivity and functional connectivity[J]. Frontiers in Neuroscience, 2018, 12: 38- 53
doi: 10.3389/fnins.2018.00038
6 MOHANTY R, SINHAA M, REMSIK A B, et al Machine learning classification to identify the stage of brain-computer interface therapy for stroke rehabilitation using functional connectivity[J]. Frontiers in Neuroscience, 2018, 12: 353- 366
doi: 10.3389/fnins.2018.00353
7 SHEN H, WANG L, LIU Y, et al Discriminative analysis of resting-state functional connectivity patterns of schizophrenia using low dimensional embedding of fMRI[J]. Neuroimage, 2010, 49 (4): 3110- 3121
doi: 10.1016/j.neuroimage.2009.11.011
8 HAMDI S M, AYDIN B, BOUBRAHIMI S F, et al. Biomarker detection from fMRI-based complete functional connectivity networks [C]// IEEE First International Conference on Artificial Intelligence and Knowledge Engineering. Laguna Hills: IEEE, 2018: 17-24.
9 CHENG W, ROLLS E T, GU H, et al Autism: reduced connectivity between cortical areas involved in face expression, theory of mind, and the sense of self[J]. Brain, 2015, 138 (5): 1382- 1393
doi: 10.1093/brain/awv051
10 KONG Y, GAO J, XU Y, et al Classification of autism spectrum disorder by combining brain connectivity and deep neural network classifier[J]. Neurocomputing, 2019, 324: 63- 68
doi: 10.1016/j.neucom.2018.04.080
11 PAWLAK Z Rough sets[J]. International Journal of Computer and Information Sciences, 1982, 11: 341- 356
doi: 10.1007/BF01001956
12 付志耀, 高岭, 孙骞, 等 基于粗糙集的漏洞属性约简及严重性评估[J]. 计算机研究与发展, 2016, 5305: 1009- 1017
FU Zhi-yao, GAO Ling, SUN Qian, et al Vulnerability attribute reduction and severity evaluation based on rough set[J]. Journal of Computer Research and Development, 2016, 5305: 1009- 1017
13 LIU H B, ABRANAM A, ZHANG W S A swarm-based rough set approach for fMRI data analysis[J]. International Journal of Innovative Computing, Information and Control, 2011, 7 (6): 3121- 3132
14 SARKHEYLI A, SOFFKER D. Automatic selection of relevant features using Rough Set Theory for real-time situation recognition based on fuzzy SOM-based CBR [C]// IEEE International Conference on Systems, Man, and Cybernetics. Budapest: IEEE, 2016: 832-837.
15 陈俞, 赵素云, 李雪峰, 等 基于随机抽样的模糊粗糙约简[J]. 软件学报, 2017, 28 (11): 2825- 2835
CHEN Yu, ZHAO Su-yun, LI Xue-feng, et al Fuzzy rough reduction based on random sampling[J]. Journal of Software, 2017, 28 (11): 2825- 2835
16 于洪, 王国胤, 姚一豫 决策粗糙集理论研究现状与展望[J]. 计算机学报, 2015, 3808: 1628- 1639
YU Hong, WANG Guo-yin, YAO Yi-yu Current situation and prospect of decision rough set theory[J]. Journal of Computer, 2015, 3808: 1628- 1639
17 HU Q H, YU D R Numerical attribute reduction based on neighborhood granulation and rough approximation[J]. Journal of Software, 2008, 19 (3): 640- 649
doi: 10.3724/SP.J.1001.2008.00640
18 MENG J, ZHANG J, LI R Gene selection using rough set based on neighborhood for the analysis of plant stress response[J]. Applied Soft Computing, 2014, 25: 51- 63
doi: 10.1016/j.asoc.2014.09.013
19 XIE X, XIE G, XU X High precision image segmentation algorithm using SLIC and neighborhood rough set[J]. Multimedia Tools and Applications, 2018, 77 (24): 31525- 31543
doi: 10.1007/s11042-018-6150-y
20 HU Q, YU D, XIE Z Neighborhood classifiers[J]. Expert Systems with Applications, 2008, 34 (2): 866- 876
doi: 10.1016/j.eswa.2006.10.043
21 李晓磊, 邵之江, 钱积新 一种基于动物自治体的寻优模式: 鱼群算法[J]. 系统工程理论与实践, 2002, 22 (11): 32- 38
LI Xiao-lei, SHAO Zhi-jiang, Qian Ji-xin A fish swarm algorithm based on animal autonomy[J]. Systems Engineering Theory and Practice, 2002, 22 (11): 32- 38
22 XIAN S, ZHANG J, XIAO Y, et al A novel fuzzy time series forecasting method based on the improved artificial fish swarm optimization algorithm[J]. Soft Computing, 2017, 22 (12): 3907- 3917
23 ZHU X, NI Z, CHENG M, et al Selective ensemble based on extreme learning machine and improved discrete artificial fish swarm algorithm for haze forecast[J]. Applied Intelligence, 2018, 48 (7): 1757- 1775
doi: 10.1007/s10489-017-1027-8
24 JIE N F, OSUCH E A, ZHU M H, et al Discriminating bipolar disorder from major depression using whole-brain functional connectivity: a feature selection analysis with SVM-FoBa algorithm[J]. Journal of Signal Processing Systems, 2018, 90 (2): 259- 271
doi: 10.1007/s11265-016-1159-9
25 JIE N F, ZHU M H, MA X Y, et al Discriminating bipolar disorder from major depression based on SVM-FoBa: efficient feature selection with multimodal brain imaging data[J]. IEEE Transactions on Autonomous Mental Development, 2015, 7 (4): 320- 331
doi: 10.1109/TAMD.2015.2440298
26 BUCKNER R L, ANDREWS-HANNA J R, SCHACTERD L The brain's default network: anatomy, function, and relevance to disease[J]. New York Academy of Sciences, 2008, 1124: 1- 38
doi: 10.1196/annals.1440.011
[1] 郭童,林峰. 基于混合遗传鱼群算法的贝叶斯网络结构学习[J]. J4, 2014, 48(1): 130-135.