基于核统计独立性准则的特征选择研究综述

doi:10.3778/j.issn.1002-8331.2203-0527

摘要/Abstract

摘要： 希尔伯特-施密特独立性准则（Hilbert-Schmidt independence criterion，HSIC）是一种基于核函数的独立性度量标准，具有计算简单、收敛速度快和偏差低等优点，广泛应用于统计分析和机器学习问题中。特征选择是一种有效的降维技术，它能评估特征的重要性，并构造适合学习任务的最优特征子空间。系统综述了基于HSIC的特征选择方法，详细介绍了其中的理论基础、算法模型和求解方法，分析了基于HSIC的特征选择的优点与不足，并对未来的研究做出展望。

关键词: 特征选择, 希尔伯特-施密特独立性准则, 核方法, 机器学习

Abstract: The Hilbert-Schmidt independence criterion（HSIC） is a kernel function-based independence criterion with the advantages of simple computation, fast convergence and low bias, which is widely used in statistical analysis and machine learning problems. Feature selection is an effective dimensionality reduction technique that evaluates the importance of features and constructs an optimal feature subspace suitable for the learning task. The HSIC-based feature selection methods are systematically reviewed, in which the theoretical basis, algorithmic models and solution methods are introduced in detail, the advantages and disadvantages of HSIC-based feature selection are analyzed, and future directions are given.

Key words: feature selection, Hilbert-Schmidt independence criterion（HSIC）, kernel method, machine learning

胡振威, 汪廷华, 周慧颖. 基于核统计独立性准则的特征选择研究综述[J]. 计算机工程与应用, 2022, 58(22): 54-64.

HU Zhenwei, WANG Tinghua, ZHOU Huiying. Review of Feature Selection Methods Based on Kernel Statistical Independence Criteria[J]. Computer Engineering and Applications, 2022, 58(22): 54-64.

参考文献

[1] KOTSIANTIS S.Feature selection for machine learning classification problems：a recent overview[J].Artificial Intelligence Review，2011，42（1）：157-176.
[2] 周志华.机器学习[M].北京：清华大学出版社，2016.
ZHOU Z H.Machine learning[M].Beijing：Tsinghua University Press，2016.
[3] JORDAN M，MITCHELL T.Machine learning：trends，perspectives，and prospects[J].Science，2015，349（6245）：255-260.
[4] HAN J，PEI J，KAMBER M.Data mining：concepts and techniques[M].Amsterdam：Elsevier，2011.
[5] ATLURI G，KARPATNE A，KUMAR V.Spatio-temporal data mining：a survey of problems and methods[J].ACM Computing Surveys，2018，51（4）：1-41.
[6] INDYK P，MOTWANI R.Approximate nearest neighbors：towards removing the curse of dimensionality[C]//30th Annual ACM Symposium on Theory of Computing，1998：604-613.
[7] CAI J，LUO J，WANG S，et al.Feature selection in machine learning：a new perspective[J].Neurocomputing，2018，300：70-79.
[8] LI J，CHENG K，WANG S，et al.Feature selection：a data perspective[J].ACM Computing Surveys，2017，50（6）：1-45.
[9] HUANG R，WU Z.Multi-label feature selection via mani-
fold regularization and dependence maximization[J].Pattern Recognition，2021，120：108149.
[10] PONTE P，MELKO R.Kernel methods for interpretable machine learning of order parameters[J].Physical Review B，2017，96（20）：205146.
[11] SHAWE-TAYLOR J，CRISTIANINI N.Kernel methods for pattern analysis[M].Cambridge：Cambridge University Press，2004.
[12] HOFMANN T，SCH?LKOPF B，SMOLA A.Kernel methods in machine learning[J].The Annals of Statistics，2008，36（3）：1171-1220.
[13] GRETTON A，HERBRICH.R，SMOLA A，et al.Kernel methods for measuring independence[J].Journal of Machine Learning Research，2005，6：2075-2129.
[14] SIMON-GABRIEL C，SCH?LKOPF B.Kernel distribution embeddings：universal kernels，characteristic kernels and kernel metrics on distributions[J].Journal of Machine Learning Research，2018，19（1）：1708-1736.
[15] 王博，黄九鸣，贾焰，等.适用于多种监督模型的特征选择方法研究[J].计算机研究与发展，2010，47（9）：1548-1557.
WANG B，HUANG J M，JIA Y，et al.A study of feature selection methods applicable to multiple supervised models[J].Journal of Computer Research and Development，2010，47（9）：1548-1557.
[16] GRETTON A，BOUSQUET O，SMOLA A，et al.Measuring statistical dependence with Hilbert-Schmidt norms[C]//2005 International Conference on Algorithmic Learning Theory.Berlin，Heidelberg：Springer，2005：63-77.
[17] BACH F，JORDAN M.Kernel independent component analysis[J].Journal of Machine Learning Research，2002，3：1-48.
[18] GRETTON A，SMOLA A，BOUSQUET O，et al.Kernel constrained covariance for dependence measurement[C]//2005 International Workshop on Artificial Intelligence and Statistics，2005：112-119.
[19] 张晨光，张燕，张夏欢.最大规范化依赖性多标记半监督学习方法[J].自动化学报，2015，41（9）：1577-1588.
ZHANG C G，ZHANG Y，ZHANG X H.Maximum normative-dependent multilabel semi-supervised learning methods[J].Acta Automatica Sinica，2015，41（9）：1577-1588.
[20] SONG L，SMOLA A，GRETTON A，et al.A dependence maximization view of clustering[C]//24th International Conference on Machine Learning，2007：815-822.
[21] SHU X，LAI D，XU H，et al.Learning shared subspace for multi-label dimensionality reduction via dependence maximization[J].Neurocomputing，2015，168：356-364.
[22] FEI X，ZHOU S，HAN X，et al.Doubly supervised parameter transfer classifier for diagnosis of breast cancer with imbalanced ultrasound imaging modalities[J].Pattern Recognition，2021，120：108139.
[23] WANG T，DAI X，LIU Y.Learning with Hilbert-Schmidt independence criterion：a review and new perspectives[J].Knowledge-Based Systems，2021，234：107567.
[24] 姚旭，王晓丹，张玉玺，等.特征选择方法综述[J].控制与决策，2012，27（2）：161-166.
YAO X，WANG X D，ZHANG Y X，et al.Overview of feature selection methods[J].Journal of Control and Decision，2012，27（2）：161-166.
[25] LIU C，MA Q，XU J.Multi-label feature selection method combining unbiased Hilbert-Schmidt independence criterion with controlled genetic algorithm[C]//International Conference on Neural Information Processing.Cham：Springer，2018：3-14.
[26] SHEIKHPOUR R，SARRAM M，GHARAGHANI S，et al.A survey on semi-supervised feature selection methods[J].Pattern Recognition，2017，64：141-158.
[27] SOLORIO-FERNáNDEZ S，CARRASCO-OCHOA J，MARTíNEZ-
TRINIDAD J F.A review of unsupervised feature selection methods[J].Artificial Intelligence Review，2020，53（2）：907-948.
[28] 张晨光，张燕，张夏欢.从希尔伯特-施密特独立性中学习的多标签半监督学习方法[J].中国科技论文，2013，8（10）：998-1002.
ZHANG C G，ZHANG Y，ZHANG X H.A multi-label semi-supervised learning approach for learning from Hilbert-Schmidt independence[J].China Sciencepaper，2013，8（10）：998-1002.
[29] 刘涛，吴功宜，陈正.一种高效的用于文本聚类的无监督特征选择算法[J].计算机研究与发展，2005，42（3）：381-386.
LIU T，WU G Y，CHEN Z.An efficient unsupervised feature selection algorithm for text clustering[J].Journal of Computer Research and Development，2005，42（3）：381-386.
[30] LIAGHAT S，MANSOORI E.Filter-based unsupervised feature selection using Hilbert-Schmidt independence criterion[J].International Journal of Machine Learning and Cybernetics，2019，10（9）：2313-2328.
[31] LUO M，NIE F，CHANG X，et al.Adaptive unsupervised feature selection with structure regularization[J].IEEE Transactions on Neural Networks and Learning Systems，2017，29（4）：944-956.
[32] NIIJIMA S，OKUNO Y.Laplacian linear discriminant analysis approach to unsupervised feature selection[J].IEEE/ACM Transactions on Computational Biology and Bioinformatics，2008，6（4）：605-614.
[33] 徐峻岭，周毓明，陈林，等.基于互信息的无监督特征选择[J].计算机研究与发展，2012，49（2）：158-168.
XU J L，ZHOU Y M，CHEN L，et al.Unsupervised feature selection based on mutual information[J].Journal of Computer Research and Development，2012，49（2）：158-168.
[34] KURNIAWATI I，PARDEDE H.Hybrid method of information gain and particle swarm optimization for selection of features of SVM-based sentiment analysis[C]//2018 International Conference on Information Technology Systems and Innovation，2018：1-5.
[35] YANG Y，PEDERSEN J.A comparative study on feature selection in text categorization[C]//14th International Conference on Machine Learning，1997：412-420.
[36] ZHAO Z，LIU H.Spectral feature selection for supervised and unsupervised learning[C]//24th International Conference on Machine Learning，2007：1151-1157.
[37] GOH L，SONG Q，KASABOV N.A novel feature selection method to improve classification of gene expression data[C]//2nd Conference on Asia-Pacific Bioinformatics-Volume 29，2004：161-166.
[38] JIN X，XU A，BIE R，et al.Machine learning techniques and chi-square feature selection for cancer classification using SAGE gene expression profiles[C]//2006 International Workshop on Data Mining for Biomedical Applications.Berlin，Heidelberg：Springer，2006：106-115.
[39] KWAK N，CHOI C H.Input feature selection for classification problems[J].IEEE Transactions on Neural Networks，2002，13（1）：143-159.
[40] 施启军，潘峰，龙福海，等.特征选择方法研究综述[J].微电子学与计算机，2022，39（3）：1-8.
SHI Q J，PAN F，LONG F H，et al.A review of feature selection methods[J].Microelectronics & Computer，2022，39（3）：1-8.
[41] REUNANEN J.Overfitting in making comparisons between variable selection methods[J].Journal of Machine Learning Research，2003，3：1371-1382.
[42] KOHAVI R，JOHN G.Wrappers for feature subset selection[J].Artificial Intelligence，1997，97（1/2）：273-324.
[43] MALDONADO S，LóPEZ J.Dealing with high-dimensional class-imbalanced datasets：embedded feature selection for SVM classification[J].Applied Soft Computing，2018，67：94-105.
[44] WANG S，TANG J，LIU H.Embedded unsupervised feature selection[C]//29th AAAI Conference on Artificial Intelligence，2015：470-476.
[45] LIU H，ZHOU M，LIU Q.An embedded feature selection method for imbalanced data classification[J].IEEE/CAA Journal of Automatica Sinica，2019，6（3）：703-715.
[46] CHANDRASHEKAR G，SAHIN F.A survey on feature selection methods[J].Computers & Electrical Engineering，2014，40（1）：16-28.
[47] PENG H，LONG F，DING C.Feature selection based on mutual information criteria of max-dependency，max-relevance，and min-redundancy[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2005，27（8）：1226-1238.
[48] FUKUMIZU K，GRETTON A，SUN X，et al.Kernel measures of conditional dependence[C]//Advances in Neural Information Processing Systems 20，2007：489-496.
[49] ZHANG X，SONG L，GRETTON A，et al.Kernel measures of independence for non-iid data[C]//Advances in Neural Information Processing Systems 21，2008：1937-1944.
[50] SONG L，SMOLA A，GRETTON A，et al.Feature selection via dependence maximization[J].Journal of Machine Learning Research，2012，13（5）：1393-1434.
[51] HERMES L，BUHMANN J.Feature selection for support vector machines[C]//15th IEEE International Conference on Pattern Recognition，2000，2：712-715.
[52] BRADLEY P，MANGASARIAN O.Feature selection via concave minimization and support vector machines[C]//15th International Conference on Machine Learning，1998：82-90.
[53] WESTON J，ELISSEEFF A，SCH?LKOPF B，et al.Use of the zero norm with linear models and kernel methods[J].Journal of Machine Learning Research，2003，3：1439-1461.
[54] HE D，RISH I，PARIDA L.Transductive HSIC Lasso[C]//2014 International Conference on Data Mining，2014：154-162.
[55] SONG L，BEDO J，BORGWARDT K，et al.Gene selection via the BAHSIC family of algorithms[J].Bioinformatics，2007，23（13）：490-498.
[56] SONG L，SMOLA A，GRETTON A，et al.Supervised feature selection via dependence estimation[C]//24th International Conference on Machine Learning，2007：823-830.
[57] WESTON J，SCHLKOPF B，ESKIN E，et al.Dealing with large diagonals in kernel matrices[J].Annals of the Institute of Statistical Mathematics，2003，55（2）：391-408.
[58] SINAGA K，YANG M.Unsupervised K-means clustering algorithm[J].IEEE Access，2020，8：80716-80727.
[59] NAKARIYAKUL S，CASASENT D.An improvement on floating search algorithms for feature subset selection[J].Pattern Recognition，2009，42（9）：1932-1940.
[60] STREARNS S.On selecting features for pattern classifiers[C]//1976 International Conference on Pattern Recognition，1976：71-75.
[61] SOMOL P，PUDIL P，NOVOVI?OVá J，et al.Adaptive floating search methods in feature selection[J].Pattern Recognition Letters，1999，20：1157-1163.
[62] LU K，SUN W，MA C，et al.Load forecast method of electric vehicle charging station using SVR based on GA-PSO[C]//IOP Conference Series：Earth and Environmental Science.2017，69：012196.
[63] YANG X.A new metaheuristic bat-inspired algorithm[J].Nature Inspired Cooperative Strategies for Optimization，2010，284：65-74.
[64] LOPEZ R，REGIER J，JORDAN M I，et al.Information constraints on auto-encoding variational Bayes[C]//Advances in Neural Information Processing Systems 31，2018：6117-6128.
[65] GANGEH M J，ZARKOOB H，GHODSI A.Fast and scalable feature selection for gene expression data using Hilbert-Schmidt independence criterion[J].IEEE/ACM Transactions on Computational Biology and Bioinformatics，2017，14（1）：167-181.
[66] MASAELI M，FUNG G，DY J.From transformation-based dimensionality reduction to feature selection[C]//27th International Conference on Machine Learning，2010.
[67] TIBSHIRANI R.Regression shrinkage and selection via the Lasso：a retrospective[J].Journal of the Royal Statistical Society：Series B（Statistical Methodology），2011，73（3）：267-288.
[68] DONOHO D.For most large underdetermined systems of linear equations the minimal [?1]-norm solution is also the sparsest solution[J].Communications on Pure and Applied Mathematics，2006，59（6）：797-829.
[69] YAMADA M，JITKRITTUM W，SIGAL L，et al.High-dimensional feature selection by feature-wise kernelized Lasso[J].Neural Computation，2014，26（1）：185-207.
[70] 李军怀，付静飞，蒋文杰，等.基于mRMR的文本分类特征选择方法[J].计算机科学，2016，43（10）：225-228.
LI J H，FU J F，JIANG W J，et al.An mRMR-based feature selection method for text classification[J].Computer Science，2016，43（10）：225-228.
[71] REN W，LI B，HAN M.A novel Granger causality method based on HSIC-Lasso for revealing nonlinear relationship between multivariate time series[J].Physica A：Statistical Mechanics and its Applications，2020，541：123245.
[72] YAMADA M，TANG J，LUGO-MARTINEZ J，et al.Ultra high-dimensional nonlinear feature selection for big biological data[J].IEEE Transactions on Knowledge and Data Engineering，2018，30（7）：1352-1365.
[73] DAMODARAN B，COURTY N，LEFèVRE S.Sparse Hilbert-Schmidt independence criterion and surrogate-kernel-based feature selection for hyperspectral image classification[J].IEEE Transactions on Geoscience and Remote Sensing，2017，55（4）：2385-2398.
[74] POIGNARD B，YAMADA M.Sparse Hilbert-Schmidt independence criterion regression[C]//23rd International Conference on Artificial Intelligence and Statistics，2020：538-548.
[75] FRANK L，FRIEDMAN J.A statistical view of some chemometrics regression tools[J].Technometrics，1993，35（2）：109-135.
[76] FAN J，LI R.Variable selection via nonconcave penalized likelihood and its oracle properties[J].Journal of the American Statistical Association，2001，96（456）：1348-1360.
[77] ZHANG C.Nearly unbiased variable selection under minimax concave penalty[J].Annals of Statistics，2010，38（2）：894-942.
[78] YAMADA M，KIMURA A，NAYA F，et al.Change-point detection with feature selection in high-dimensional time-series data[C]//23rd International Joint Conference on Artificial Intelligence，2013.
[79] KONG X，PHILIP S.gMLC：a multi-label feature selection framework for graph classification[J].Knowledge and Information Systems，2012，31（2）：281-305.
[80] CAMPS-VALLS G，MOOIJ J，SCHOLKOPF B.Remote sensing feature selection by kernel dependence measures[J].IEEE Geoscience and Remote Sensing Letters，2010，7（3）：587-591.
[81] XU J.Effective and efficient multi-label feature selection approaches via modifying Hilbert-Schmidt independence criterion[C]//23rd International Conference on Neural Information Processing.Cham：Springer，2016：385-395.
[82] JIANG L，WANG J，YU G.Semi-supervised multi-label feature selection based on sparsity regularization and dependence maximization[C]//2018 9th International Conference on Intelligent Control and Information Processing，2018：325-332.
[83] LIU Y，ZHANG C，ZHU P，et al.Generalized multi-view unsupervised feature selection[C]//27th International Conference on Artificial Neural Networks.Cham：Springer，2018：469-478.