计算机科学

计算机科学 ›› 2023, Vol. 50 ›› Issue (6): 167-174.doi: 10.11896/jsjkx.220900144

• 数据库&大数据&数据科学 • 上一篇    下一篇

基于超图正则化的多模态信息融合算法

崔冰晶, 张懿璞, 王飚   

  1. 长安大学电子与控制工程学院 西安 710061
  • 收稿日期:2022-09-15 修回日期:2022-12-09 出版日期:2023-06-15 发布日期:2023-06-06
  • 通讯作者: 张懿璞(zyipu@chd.edu.cn)
  • 作者简介:(2020132049@chd.edu.cn)
  • 基金资助:
    国家重点研发计划(2021YFB1600200,2021YFB2601300)

Multimodal Data Fusion Algorithm Based on Hypergraph Regularization

CUI Bingjing, ZHANG Yipu, WANG Biao   

  1. School of Electronic and Control Engineering,Chang'an University,Xi'an 710061,China
  • Received:2022-09-15 Revised:2022-12-09 Online:2023-06-15 Published:2023-06-06
  • About author:CUI Bingjing,born in 1996,master candidate.Her main research interests include multi-modal data fusion and machine learning.WANG Biao,born in 1969,Ph.D,professor.His main research interests include analysis and optimization of complex networks and systems,multi-agent control.
  • Supported by:
    National Key R & D Program of China(2021YFB1600200,2021YFB2601300).

PDF (PC)

328

摘要： 多模态数据融合方法通过学习多个数据集间的关联信息和互补信息,提高了数据分类或预测的性能。但现有的数据融合方法大都基于单独数据集自身的特征模式进行学习,不同异构数据之间的结构信息往往被忽略。因此,文中提出了一种基于超图正则化的多模态信息融合算法(sHMF),通过超图和流行正则项的方法结合表示模态内样本间的高阶关系和模态间的关系,即得到同构和异构的高阶网络。其中,采用超图稀疏表达学习超图,减少冗余边。为了验证所提算法的性能,在模拟数据和影响遗传学真实数据下进行实验,结果表明,sHMF算法在模拟数据和真实数据上均优于多任务学习、多邻域分类等流行算法对精神分裂症的分类精度。同时,sHMF在真实数据上得出的实验结果进一步揭示了一些与精神分裂症显著相关的生物标记物以及风险基因、甲基化因子和异常脑区之间潜在的联系。

关键词: 多模态数据融合, 流形正则化, 功能磁共振, 影像遗传学

Abstract: The multi-modal data fusion improves the performance of data classification and prediction by learning the correlation information and complementary information between multiple datasets.However,existing data fusion methods are based on feature pattern learning of single dataset and ignore structural information among different heterogeneous datasets.This paper proposes a multi-modal data fusion algorithm based on hypergraph regularization(sHMF),acquiring hyper-order relationship of inter-and cross-modality by combining hypergraph and manifold regularization,i.e.homogeneous and heterogeneous high-order networks.Specifically,it firstly generates a hypergraph similarity matrix to represent the high-order relationships among subjects.In the proposed method,the sparse representation of hypergraph is used to build hypergraph for reducing redundant hype-redges.sHMF is validated on the simulated data and real imaging genetic data of schizophrenia patients.Experiment results show that our algorithm outperforms several widely used methods in the classification accuracy of simulated data and real data,and reveals some biomarkers significantly associated with schizophrenia and the potential links between risk genes,methylation factors and abnormal brain regions.

Key words: Multi-modal data fusion, Manifold regularization, Functional magnetic resonance imaging, Imaging genetics

中图分类号: 

  • TP391
[1]REN Z Y,WANG Z C,KE Z W,et al.Overview of multimodal data fusion [J].Computer Engineering and Application,2021,57(18):49-64.
[2]ADELI E,MENG Y,LI G,et al.Multi-task prediction of infant cognitive scores from longitudinal incomplete neuroimaging data[J].Neuroimage,2019,185:783-792.
[3]JIE B,ZHANG D,CHENG B,et al.Manifold regularized multitask feature learning for multimodality disease classification[J].Human brain mapping,2015,36(2):489-507.
[4]ZHU X,SUK H I,LEE S W,et al.Subspace regularized sparse multitask learning for multiclass neurodegenerative disease identification[J].IEEE Transactions on Biomedical Engineering,2015,63(3):607-618.
[5]DU L,LIU K,YAO X,et al.Multi-Task Sparse Canonical Correlation Analysis with Application to Multi-Modal Brain Imaging Genetics[J].IEEE/ACM Transactions on Computational Biology and Bioinformatics,2021,18(1):227-239.
[6]BAI Y,PASCAL Z,CALHOUN V,et al.Optimized Combination of Multiple Graphs with Application to the Integration of Brain Imaging and Genomics Data[J].IEEE Transactions on medical imaging,2019,39(6):1801-1811.
[7]HU W,MENG X,BAI Y,et al.Interpretable multimodal fusion networks reveal mechanisms of brain cognition[J].IEEE Transactions on medical imaging,2021,40(5):1474-1483.
[8]JI R,GAO Y,HONG R,et al.Spectral-Spatial Constraint Hyperspectral Image Classification[J].IEEE Transactions on Geoscience and Remote Sensing,2014,52(3):1811-1824.
[9]XIAO L,WANG J,KASSANI P H,et al.Multi-HypergraphLearning-Based Brain Functional Connectivity Analysis in fMRI Data[J].IEEE Transactions on Medical Imaging,2020,39(5):1746-1758.
[10]XIAO L,STEPHEN J M,WILSON T W,et al.A manifold regularized multi-task learning model for IQ prediction from two fMRI paradigms[J].IEEE Transactions on Biomedical Engineering,2019,67(3):796-806.
[11]PENG Y.Research on Multimodal Feature Selection and Classification Method Based on Hypergraph [D].Nanjing:Nanjing University of Aeronautics and Astronautics,2019.
[12]ARGYRIOU A,EVGENIOU T,PONTIL M.Multi-Task Fea-ture Learning[C]//Conference on Advances in Neural Information Processing Systems.Cambridge:MIT Press,2007:19-41.
[13]WANG B,MEZLINI A M,DEMIR F,et al.Similarity network fusion for aggregating data types on a genomic scale[J].Nat Methods,2014,11(3):333-337.
[14]DENG S P,HU W,CALHOUN V D,et al.Schizophrenia prediction using integrated imaging genomic networks[J].Advances in Science,Technology and Engineering Systems,2017,2(3):702-710.
[15]HU W,LIN D,CAO S,et al.Adaptive Sparse Multiple Canonical Correlation Analysis with Application to Imaging (Epi)Genomics Study of Schizophrenia[J].IEEE Transactions on Biomedical Engineering,2018,65(2):390-399.
[16]TZOURIO-MAZOYER N,LANDEAU B,PAPATHANAS-SIOU D,et al.Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain[J].Neuroimage,2002,15(1):273-289.
[17]SUN L,PATEL R,LIU J,et al.Mining brain region connectivity for Alzheimer's disease study via sparse inverse covariance estimation[C]//Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.Paris France:KDD09,2009:1335-1344.
[18]LIN D,CALHOUN V D,WANG Y P.Correspondence between fMRI and SNP data by group sparse canonical correlation analysis[J].Medical Image Analysis,2014,18(6):891-902.
[19]PIDSLEY R,Y WONG C C,VOLTA M,et al.A data-driven approach to preprocessing Illumina 450K methylation array data[J].BMC genomics,2013,14(1):1-10.
[20]LIU J,CHEN J,EHRLICH S,et al.Methylation patterns in whole blood correlate with symptoms in schizophrenia patients[J].Schizophrenia bulletin,2014,40(4):769-776.
[21]CHENG Y J,SUN Q Q,ZHAO M.Epigenetic research progress of methamphetamine use and addiction[J].Shanghai:Journal of Shanghai JiaoTong University (Medical Edition),2021,41(8):1094-1098.
[22]LI M Y C,ZHANG W N,XIONG X,et al.Research progress on the balance of central excitation inhibition in an autism model with SHANK3 gene mutation [J].Life Sciences,2021,33(8):962-970.
[23]FANG J,LIN D,SCHULZ C,et al.Joint sparse canonical correlation analysis for detecting differential imaging genetics mo-dules[J].Bioinformatics,2016,32(22):3480-3488.
[24]PENG P,ZHANG Y,JU Y,et al.Group Sparse Joint Non-negative Matrix Factorization on Orthogonal Subspace for Multi-modal Imaging Genetics Data Analysis[J].IEEE/ACM Transactions on Computational Biology and Bioinformatics,2022,19(1):479-490.
[25]WANG M,HUANG T Z,FANG J,et al.Integration of Imaging (Epi) Genomics Data for the Study of Schizophrenia Using Group Sparse Joint Nonnegative Matrix Factorization[J].IEEE/ACM Transactions on Computational Biology and Bioinforma-tics,2020,17(5):1671-1681.
[26]KWON E,WANG W,TSAI L H.Validation of schizophrenia-associated genes CSMD1,C10orf26,CACNA1C and TCF4 as miR-137 targets[J].Mol Psychiatry,2013,18(1):11-12.
[27]LEE M R,SHESKIER M B,FAROKHNIA M,et al.Oxytocin receptor mRNA expression in dorsolateral prefrontal cortex in major psychiatric disorders:A human post-mortem study[J].Psychoneuroendocrinology,2018,96(1):143-147.
[28]ZHANG Y,YOU X,LI S,et al.Peripheral Blood Leukocyte RNA-Seq Identifies a Set of Genes Related to Abnormal Psychomotor Behavior Characteristics in Patients with Schizophrenia[J].Medical Science Monitor:International Medical Journal of Experimental and Clinical Research,2020,26(1):1-31.
[29]SAMAAN M C.Prader-Willi Syndrome:Genetics,Phenotype,and Management[J].Current Psychiatry Reviews,2014,10(2):168-181.
[30]MARAZZITI D,BARONI S,PICCHETTI M,et al.Psychiatric disorders and mitochondrial dysfunctions[J].European Review Medical and Pharmacological Sciences,2012,16(2):270-275.
[31]GRANT A,FATHALLI F,ROULEAU G,et al.Association between schizophrenia and genetic variation in DCC:A case-control study[J].Schizophrenia Research,2012,137(1):26-31.
[32]JOB D E,WHALLEY H C,MCCONNELL S,et al.Structural gray matter differences between first-episode schizophrenics and normal controls using voxel-based morphometry[J].Neuroi-mage,2002,17(2):880-889.
[33]DUAN H F,GAN J L,YANG J M,et al.A longitudinal study on intrinsic connectivity of hippocampus associated with positive symptom in first-episode schizophrenia[J].Behavioral brain research,2015,283(1):78-86.
[34]ZHENG J J.Research on multimodal brain network of schizophrenia based on magnetic resonance imaging[D].Chengdu:University of Electronic Science and Technology,2017.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发