Model selection and estimation in the Gaussian graphical model
The DOI matches its Crossref record; 16 of the 16 references it cites were found in a registry or an index.
Abstract
We propose penalized likelihood methods for estimating the concentration matrix in the Gaussian graphical model. The methods lead to a sparse and shrinkage estimator of the concentration matrix that is positive definite, and thus conduct model selection and estimation simultaneously. The implementation of the methods is nontrivial because of the positive definite constraint on the concentration matrix, but we show that the computation can be done effectively by taking advantage of the efficient maxdet algorithm developed in convex optimization. We propose a BIC-type criterion for the selection of the tuning parameter in the penalized likelihood methods. The connection between our methods and existing methods is illustrated. Simulations and real examples demonstrate the competitive performance of the new methods.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“These methods have the advantage to be able to recover the off diagonal structure of the covariance matrix.”
From Signal Estimation and Uncertainties Extraction in Terahertz Time-Domain Spectroscopy (2025) · cites this paper as [58]
Citations
Cites
Boyd, S., & Vandenberghe, L. (2004) · doi:10.1017/cbo9780511804441 · Cited by 56,601
Regression shrinkage and selection via the lasso
Tibshirani, R. (1996) · Journal of the Royal Statistical Society Series B: Statistical Methodology · doi:10.1111/j.2517-6161.1996.tb02080.x · Cited by 51,482 · cited 1×
“This is similar to the idea of the lasso in linear regression (Tibshirani, 1996).”
MethodGraphical Models in Applied Multivariate Statistics
Goodall, C. R. (1991) · Technometrics · doi:10.1080/00401706.1991.10484880 · Cited by 1,849
Asymptotics for lasso-type estimators
Fu, W., & Knight, K. (2000) · The Annals of Statistics · doi:10.1214/aos/1015957397 · Cited by 1,276 · cited 2×
“, X(p)) be a p-dimensional random vector following a multivariate normal distribution Np(µ, ) with unknown mean µ and nonsingular covariance matrix .”
BackgroundMethodInfluentialHeuristics of instability and stabilization in model selection
Breiman, L. (1996) · The Annals of Statistics · doi:10.1214/aos/1032181158 · Cited by 1,146 · cited 1×
“As demonstrated by Breiman (1996), the discrete nature of such procedures often leads to instability of the estimator: small changes in the data may result in very different estimates.”
BackgroundBetter Subset Regression Using the Nonnegative Garrote
Breiman, L. (1995) · Technometrics · doi:10.1080/00401706.1995.10484371 · Cited by 1,043 · cited 2×
“As demonstrated by Breiman (1996), the discrete nature of such procedures often leads to instability of the estimator: small changes in the data may result in very different estimates.”
BackgroundMethodIntroduction to Graphical Modelling
Edwards, D. (2000) · Springer Texts in Statistics · doi:10.1007/978-1-4612-0493-0 · Cited by 825
Determinant Maximization with Linear Matrix Inequality Constraints
Vandenberghe, L., Boyd, S., & Wu, S.-P. (1998) · SIAM Journal on Matrix Analysis and Applications · doi:10.1137/s0895479896303430 · Cited by 652 · cited 2×
“To use the algorithm of Vandenberghe et al. (1998), it is also required that Fi, i = 1,. .”
MethodMultivariate Dependencies: Models, Analysis and Interpretation.
Murphy, S. A., Cox, D. R., & Wermouth, N. (1998) · Biometrics · doi:10.2307/2534028 · Cited by 487
Sparsistency and rates of convergence in large covariance matrix estimation
Lam, C., & Fan, J. (2009) · The Annals of Statistics · doi:10.1214/09-aos720 · Cited by 418
Cited by
Items 101–200 of 1,000
Sparse matrix inversion with scaled Lasso
Sun, T., & Zhang, C. (2013) · Journal of Machine Learning Research · Cited by 165
Fused Multiple Graphical Lasso
Yang, S., Lu, Z., Shen, X., Wonka, P., & Ye, J. (2015) · SIAM Journal on Optimization · doi:10.1137/130936397 · Cited by 114
“One of the most widely used graphical models is the Gaussian graphical model (GGM), which assumes the variables to be Gaussian distributed [2, 54].”
BackgroundMethodInfluentialSparse Matrix Graphical Models
Leng, C., & Tang, C. Y. (2012) · Journal of the American Statistical Association · doi:10.1080/01621459.2012.706133 · Cited by 124
“…a sparse graphical model for vec( X ), there exists abundant literature that makes use of penalized likelihood (Meinshausen and B¨uhlmann 2006; Yuan and Lin 2007; Banejee, El Ghaoui, and d’Aspremont 2008; Rothman et al. 2008; Fan, Feng, and Wu 2009; Lam and Fan 2009; Peng et al. 2009; Guo et…”
MethodRobust graphical modeling of gene networks using classical and alternative t-distributionsPreprint
Finegold, M., & Drton, M. (2011) · Cited by 104
Raskutti, G., Yu, B., Wainwright, M. J., & Ravikumar, P. (2008) · neural information processing systems · Cited by 147
Penalized model-based clustering with unconstrained covariance matrices
Zhou, H. M., Pan, W., & Shen, X. (2009) · Electronic Journal of Statistics · doi:10.1214/09-ejs487 · Cited by 102
“2006 [17]; Yuan and Lin 2007 [53]; Levina et al.”
BackgroundFeature-Inclusion Stochastic Search for Gaussian Graphical Models
Scott, J. G., & Carvalho, C. M. (2008) · Journal of Computational and Graphical Statistics · doi:10.1198/106186008x382683 · Cited by 115
“Several methods for choosing each regression model are available, some based upon L 1-regularization (Meinshausen and Buhlmann 2006; Yuan and Lin 2007) and others based upon stepwise selection (Dobra et al. 2004).”
MethodHigh-dimensional learning of linear causal networks via inverse covariance estimationPreprint
Loh, P., & Bühlmann, P. (2013) · arXiv (Cornell University) · doi:10.48550/arxiv.1311.3492 · Cited by 103
“Unfortunately, appropriate tools for inferring such superstructures are rather limited, and the usual method of using the graphical Lasso to estimate a conditional independence graph is rigorously justified only in the linear Gaussian setting (Yuan and Lin, 2007).”
MethodReview of Biological Network Data and Its Applications
Yu, D., Kim, M., Xiao, G., & Hwang, T. H. (2013) · Genomics & Informatics · doi:10.5808/gi.2013.11.4.200 · Cited by 99
“These methods can be categorized into four types: regression-based [31, 32], penalized likelihood [33-38], empirical Bayes [39], and constrained l1 minimization [40].”
BackgroundMethodThe sparse Laplacian shrinkage estimator for high-dimensional regressionPreprint
Huang, J., Ma, S., & Li, H. (2010) · Cited by 100
Fundamental Performance Limits for Ideal Decoders in High-Dimensional Linear Inverse Problems
Bourrier, A., Davies, M. E., Peleg, T., Pérez, P., & Gribonval, R. (2014) · IEEE Transactions on Information Theory · doi:10.1109/tit.2014.2364403 · Cited by 99
Zhao, Q., Sang, X., Metmer, H., Swati, Z. N. K., & Lu, J. (2019) · Cortex · doi:10.1016/j.cortex.2019.04.026 · Cited by 102
“…many methods in the past decade have been proposed based on the sample covariance estimator, e.g., Banerjee, Ghaoui, & d ' Aspremont (2008), Yuan & Lin (2007) and Friedman, Hastie, & Tibshirani (2008) take advantage of the Gaussian likelihood proposed the graphic lasso (GLASSO) estimator…”
Ridge estimation of inverse covariance matrices from high-dimensional data
van Wieringen, W. N., & Peeters, C. F. W. (2016) · Computational Statistics & Data Analysis · doi:10.1016/j.csda.2016.05.012 · Cited by 90
“rtial correlation coecients (indexed by, say, j<j0) follow a mixture distribution: f n [P^Ia( a)] jj0 o = 0f 0 n [P^Ia( a)] jj0; o + (1 0)f E n [P^Ia( a)] jj0 o ; with mixture weight 0 2[0;1], and where f 0fgdenotes the distribution of a null-edge while f Efgdenotes the distribution of a present edge. The former de…”
BackgroundGaussian and Mixed Graphical Models as (multi-)omics data analysis tools
Altenbuchinger, M., Weihs, A., Quackenbush, J. F., Grabe, H. J., & Zacharias, H. U. (2019) · Biochimica et Biophysica Acta (BBA) - Gene Regulatory Mechanisms · doi:10.1016/j.bbagrm.2019.194418 · Cited by 106
“…model to perform neighborhood selection for all nodes simultaneously (SPACE) (Peng et al., 2009), a penalized maximum likelihood approach (Yuan and Lin, 2007; Banerjee et al., 2008; Friedman et al., 2008), and a bivariate nodewise-scaled LASSO method called asymptotic normal thresholding…”
MethodFast and adaptive sparse precision matrix estimation in high dimensions
Liu, W., & Luo, X. (2014) · Journal of Multivariate Analysis · doi:10.1016/j.jmva.2014.11.005 · Cited by 112
“In particular, the ℓ 1 penalized normal likelihood estimator and its variants, which shall be called ℓ 1 -MLE estimators, were considered in several papers; see, for example, Yuan and Lin (2007), Friedman et al. (2008), Banerjee et al. (2008), and Rothman et al. (2008).”
MethodTemporal hierarchies with autocorrelation for load forecasting
Nystrup, P., Lindström, E., Pinson, P., & Madsen, H. (2019) · European Journal of Operational Research · doi:10.1016/j.ejor.2019.07.061 · Cited by 103
“With the series variances the estimator is where the inverse cross-correlation matrix Θ is found by maximizing the penalized Gaussian log-likelihood We follow Yuan and Lin (2007) by omitting the diagonal elements from the penalty.”
MethodJoint Estimation of Multiple Graphical Models from High Dimensional Time Series
Qiu, H., Han, F., Liu, H., & Caffo, B. S. (2015) · Journal of the Royal Statistical Society Series B (Statistical Methodology) · doi:10.1111/rssb.12123 · Cited by 116
“res. In settings where the dimension, d, is much larger than the sample size, n , Meinshausen and Buhlmann (2006) proposed to estimate G by solving a collection of regression problems via the lasso. Yuan and Lin (2007), Banerjee et al. (2008), Friedman et al. (2008), Rothman et al. (2008), and Liu and Luo (2012) propo…”
Missing values: sparse inverse covariance estimation and an extension to sparse regression
Städler, N., & Bühlmann, P. (2010) · Statistics and Computing · doi:10.1007/s11222-010-9219-7 · Cited by 116
“Here (μ̂, K̂) denotes the MissGLasso estimator (8) using the tuning parameter λ and df =∑j≤j ′ 1{K̂jj ′ =0} are the degrees of freedom (Yuan and Lin 2007).”
InfluentialRobust graphical modeling of gene networks using classical and alternative t-distributions
Finegold, M., & Drton, M. (2011) · The Annals of Applied Statistics · doi:10.1214/10-aoas410 · Cited by 136
Slawski, M., & Hein, M. (2014) · Linear Algebra and its Applications · doi:10.1016/j.laa.2014.04.020 · Cited by 115
“In the references [55, 3, 22, 44, 21, 42] l1-regularized log-determinant divergence minimization, which amounts to l1-penalized maximum likelihood estimation in the Gaussian case, is investigated.”
BackgroundModel selection and estimation in the matrix normal graphical model
Yin, J., & Li, H. (2012) · Journal of Multivariate Analysis · doi:10.1016/j.jmva.2012.01.005 · Cited by 103
“Similar techniques can be applied to derive Using the same argument as in [5], we have”
BackgroundMethodResultInfluentialNetwork-based approaches for analysis of complex biological systems
Chasman, D. A., Siahpirani, A. F., & Roy, S. (2016) · Current Opinion in Biotechnology · doi:10.1016/j.copbio.2016.04.007 · Cited by 93
The sparse Laplacian shrinkage estimator for high-dimensional regression
Huang, J., Ma, S., Li, H., & Zhang, C.-H. (2011) · The Annals of Statistics · doi:10.1214/11-aos897 · Cited by 118
“See, for example, Zhang and Horvath (2005), Chung and Lu (2006), Meinshausen and Bühlmann (2006), Yuan and Lin (2007), Friedman, Hastie and Tibshirani (2008), Fan, Feng and Wu (2009), among others.”
BackgroundChatterjee, A., & Lahiri, S. N. (2013) · The Annals of Statistics · doi:10.1214/13-aos1106 · Cited by 110
“Other popular penalized estimation and variable selection methods are given by the SCAD [Fan and Li (2001)] and the Dantzig Selector [Can-des and Tao (2007)].”
BackgroundInfluentialSmooth Optimization Approach for Sparse Covariance Selection
Lu, Z. (2009) · SIAM Journal on Optimization · doi:10.1137/070695915 · Cited by 105
“As shown in d’Aspremont et al. [9] (see also [3]), and Yuan and Lin [25], the l1-norm penalized maximum likelihood estimation problems (16) and (17) are capable of discovering effectively the sparse structure, or equivalently, the conditional independence in the underlying graphical model.”
BackgroundMethodInfluentialOn Convergence of Kronecker Graphical Lasso Algorithms
Tsiligkaridis, T., Hero, A. O., & Zhou, S. (2013) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2013.2240157 · Cited by 89
“For example, the Glasso model of Yuan and Lin [2] and Banerjee et al [3] imposes sparsity constraints on the covariance.”
MethodInfluentialLearning Graphical Models With Hubs
Tan, K. M., London, P., Mohan, K., Lee, S., Fazel, M., & Witten, D. (2014) · Journal of Machine Learning Research · doi:10.48550/arxiv.1402.7349 · Cited by 114
Computational Aspects Related to Inference in Gaussian Graphical Models With the G-Wishart Prior
Lenkoski, A., & Dobra, A. (2010) · Journal of Computational and Graphical Statistics · doi:10.1198/jcgs.2010.08181 · Cited by 112
“Among many notable contributions focused on structural learning in Gaussian graphical models we mention the regularization methods of Yuan and Lin (2007), Bickel and Levina (2008), Meinshausen and Bühlmann (2006).”
MethodA Local Poisson Graphical Model for Inferring Networks From Sequencing Data
Allen, G. I., & Liu, Z. (2013) · IEEE Transactions on NanoBioscience · doi:10.1109/tnb.2013.2263838 · Cited by 110
“Many have developed methods to estimate high-dimensional Markov Networks for Gaussian or binary data by using sparsity to select the edges between genes [4], [5], [6], [7].”
MethodSpatiotemporal Models for Data-Anomaly Detection in Dynamic Environmental Monitoring Campaigns
Dereszynski, E. W., & Dietterich, T. G. (2011) · ACM Transactions on Sensor Networks · doi:10.1145/1993042.1993045 · Cited by 86
Bayesian structure learning in graphical models
Banerjee, S., & Ghosal, S. (2015) · Journal of Multivariate Analysis · doi:10.1016/j.jmva.2015.01.015 · Cited by 114
“These include banding, thresholding, tapering and penalization based methods; for example, see Ledoit and Wolf (2004); Huang et al. (2006); Yuan and Lin (2007); Bickel and Levina (2008a,b); Karoui (2008); Friedman et al. (2008); Rothman et al. (2008); Lam and Fan (2009); Rothman et al. (2009); Cai…”
MethodInfluentialTopology Selection in Graphical Models of Autoregressive Processes
Songsiri, J., & Vandenberghe, L. (2010) · Journal of Machine Learning Research · doi:10.5555/1756006.1953020 · Cited by 112
“Meinshausen and B¨uhlmann [MB06] give explicit formulas for γ based on a statistical analysis of the probability of errors in the topology (see also [ YL07 , BEd08]).”
BackgroundInfluentialXie, B., Pan, W., & Shen, X. (2008) · Electronic Journal of Statistics · doi:10.1214/08-ejs194 · Cited by 104
“is how to guarantee a shrunken covariance matrix to be positive definite, as evidenced by the challenge in a simpler context of penalized estimation of a single covariance matrix (Huang et 33 al 2005; Yuan and Lin 2007). An alternative approach is to have a model intermediate between the independent and unrestricted mo…”
BackgroundMethodHigh-dimensional covariance estimation based on Gaussian graphical modelsPreprint
Zhou, S., Rütimann, P., Xu, M., & Bühlmann, P. (2010) · arXiv (Cornell University) · doi:10.48550/arxiv.1009.0530 · Cited by 87
“…matrix Σ − 1 is given by the ℓ 1 -norm regularized maximum-likelihood estimation, also known as the GLasso [Banerjee et al., 2008; Friedman et al., 2007; Yuan and Lin, 2007]. b The GLasso approach is simple to use, at least when relying on publicly available software such as the glasso package in…”
BackgroundMethodInfluentialInference via low-dimensional couplingsPreprint
Spantini, A., Bigoni, D., & Marzouk, Y. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1703.06131 · Cited by 84
“rkov structure. In other settings, the graph is unknown and must be estimated. When only samples from ˇ are available, this is a question of model learning, as described in [62, Part III]; see also [51,74,124,67] for various applications. In case of a known and smooth target density, we can characterize pairwise condi…”
MethodEfficient Distributed Learning with SparsityPreprint
Wang, J., Kolar, M., Srebro, N., & Zhang, T. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1605.07991 · Cited by 83
“Widely studied graphical models are Gaussian graphical models (Meinshausen and Bühlmann, 2006; Yuan and Lin, 2007) for continuous data and Ising graphical models (Ravikumar et al.”
BackgroundAlternating Direction Methods for Latent Variable Gaussian Graphical Model Selection
Ma, S., Xue, L., & Zou, H. (2013) · Neural Computation · doi:10.1162/neco_a_00379 · Cited by 99
“The l1-penalized maximum normal likelihood method is another popular method for graphical model selection [59, 2, 21, 47].”
MethodAn alternative to post hoc model modification in confirmatory factor analysis: The Bayesian lasso.
Pan, J., Ip, E. H., & Dubé, L. (2017) · Psychological Methods · doi:10.1037/met0000112 · Cited by 85
“One example is the graphical Lasso (Yuan & Lin, 2007; Friedman et al., 2007) for fitting a sparse Gaussian graph that applied the L1-penalty — i.e., the sum of |θj |, to the inverse covariance matrix for graph-edge selection.”
Integrative network analysis of TCGA data for ovarian cancer
Zhang, Q., Burdette, J. E., & Wang, J.-P. (2014) · BMC Systems Biology · doi:10.1186/s12918-014-0136-9 · Cited by 85
“We say The GBN is the popular BN model and its structure learning problem has been discussed by several researchers [20,38].”
BackgroundMethodData Analytics on Graphs Part III: Machine Learning on Graphs, from Graph Topology to Applications
Stanković, L., Mandic, D. P., Daković, M., Brajović, M., Dees, B. S., Li, S., & Constantinides, A. G. (2020) · Foundations and Trends® in Machine Learning · doi:10.1561/2200000078-3 · Cited by 99
Improving Mean Variance Optimization through Sparse Hedging Restrictions
Goto, S., & Xu, Y. (2015) · Journal of Financial and Quantitative Analysis · doi:10.1017/s0022109015000526 · Cited by 95
“To promote shrinkage and subset selection (sparsity), we estimate Ψ by quasi-maximum likelihood with a constraint on the sum of the absolute values (i.e., l1 norm) of its off-diagonal elements, similarly to Yuan and Lin (2007) and Rothman, Bickel, Levina, and Zhu (2008).”
MethodAlternating Direction Method for Covariance Selection Models
Yuan, X. (2011) · Journal of Scientific Computing · doi:10.1007/s10915-011-9507-1 · Cited by 80
“To perform the covariance selection effectively, it was proposed in [5] (also in [35] for a slightly different variant) to maximize the log-likelihood penalized by the l1 norm: max log(det(X)) − 〈 ,X〉 − ρe |X|e (1.”
MethodInferring multiple graphical structures
Chiquet, J., Grandvalet, Y., & Ambroise, C. (2010) · Statistics and Computing · doi:10.1007/s11222-010-9191-2 · Cited by 97
“…parameters can be guided by theoretical results that provide a bound on the rate of false edge discovery (Mein-shausen and B¨uhlmann 2006, Banerjee et al. 2008, Ambroise et al. 2009), or by more traditional information criteria targeting the estimation of K (Yuan and Lin 2007, Rocha et al. 2008).”
BackgroundSparse estimation via nonconcave penalized likelihood in factor analysis model
Hirose, K., & Yamamoto, M. (2014) · Statistics and Computing · doi:10.1007/s11222-014-9458-0 · Cited by 95
Fu, F., & Zhou, Q. (2012) · Journal of the American Statistical Association · doi:10.1080/01621459.2012.754359 · Cited by 103
“Yuan and Lin (2007) proposed to maximize an L 1 -penalized log-likelihood based on the “max-det” problem considered by Vandenberghe, Boyd, and Wu (1998), while Banerjee, El Ghaoui, and d’Aspremont (2008) employed a blockwise coordinate descent (CD) algorithm to solve the optimization problem.”
MethodLearning Brain Connectivity of Alzheimer's Disease from Neuroimaging Data
Huang, S., Li, J., Sun, L., Liu, J., Wu, T., Chen, K., Fleisher, A., Reiman, E. M., & Ye, J. (2009) · neural information processing systems · Cited by 106
Simultaneous Clustering and Estimation of Heterogeneous Graphical Models
Hao, B., Sun, W. W., Liu, Y., & Cheng, G. (2016) · PubMed · doi:10.48550/arxiv.1611.09391 · Cited by 78
“For instance, in online advertising, an important task is to find the most suitable advertisement (ad) for a given user in a specific online context.”
FastGGM: An Efficient Algorithm for the Inference of Gaussian Graphical Model in Biological Networks
Wang, T., Ren, Z., Ding, Y., Fang, Z., Sun, Z., MacDonald, M. L., Sweet, R. A., Wang, J., & Chen, W. (2016) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1004755 · Cited by 83
Network Reconstruction From High-Dimensional Ordinary Differential Equations
Chen, S., Shojaie, A., & Witten, D. (2017) · Journal of the American Statistical Association · doi:10.1080/01621459.2016.1229197 · Cited by 96
“so (Tibshirani, 1996) and the group lasso (Yuan and Lin, 2006) have been well-studied. Such penalties have also been extensively used to recover the structure of probabilistic graphical models (e.g., Yuan and Lin, 2007; Friedman et al., 2008; Meinshausen and Buhlmann, 2010; Voorman et al., 2014). However,¨ model selec…”
MethodThe Graphical Horseshoe Estimator for Inverse Covariance Matrices
Li, Y., Craig, B. Α., & Bhadra, A. (2019) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2019.1575744 · Cited by 106
“|) in Equation (1) can be taken with or without a penalty on the diagonal terms (Rothman et al. 2008; Meinshausen and Bühlmann 2006; Yuan and Lin 2007; Friedman, Hastie, and Tibshirani 2008).”
BackgroundAlternating Proximal Gradient Method for Convex Minimization
Ma, S. (2015) · Journal of Scientific Computing · doi:10.1007/s10915-015-0150-0 · Cited by 81
“…type arise from compressed sensing [9,20] and its extensions and applications in sparse MRI [53], low-rank matrix completion [8,10,44,63], sparse data fitting [67], sparse principal component analysis [17,81], sparse graphical model selection [2,26,77], robust principal component pursuit [7,12] etc.”
BackgroundLearning Joint Semantic Parsers from Disjoint Data
Peng, H., Thomson, S., Swayamdipta, S., & Smith, N. A. (2018) · Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) · doi:10.18653/v1/n18-1135 · Cited by 85
“The above technique is closely related to a line of work in estimating the structure of sparse graphical models (Yuan and Lin, 2007; Friedman et al., 2008), where an `1 penalty is applied to the inverse covariance matrix in order to induce a smaller number of conditional dependencies between variables.”
MethodHigh-Dimensional Mixed Graphical Models
Cheng, J., Li, T., Levina, E., & Zhu, J. (2016) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2016.1237362 · Cited by 89
High-dimensional analysis of semidefinite relaxations for sparse principal components
Amini, A., & Wainwright, M. J. (2008) · IEEE International Symposium on Information Theory (ISIT) · doi:10.1109/isit.2008.4595432 · Cited by 135
Sparse Estimation of Conditional Graphical Models With Application to Gene Networks
Li, B., Chun, H., & Zhao, H. (2012) · Journal of the American Statistical Association · doi:10.1080/01621459.2011.644498 · Cited by 81
“The proof of part 3 is similar to that of Theorem 1 in Yuan and Lin (2007) in the context of GGM.”
BackgroundMethodInfluentialBayesian inference of network structure from unreliable data
Young, J.-G., Cantwell, G. T., & Newman, M. E. J. (2020) · Journal of Complex Networks · doi:10.1093/comnet/cnaa046 · Cited by 91
Nemesure, M. D., Collins, A. C., Price, G. D., Griffin, T. Z., Pillai, A., Nepal, S. K., Heinz, M. V., Lekkas, D., Campbell, A. T., & Jacobson, N. C. (2024) · Journal of Psychopathology and Clinical Science · doi:10.1037/abn0000884 · Cited by 79
On the distribution of the adaptive LASSO estimator
Pötscher, B. M., & Schneider, U. (2009) · Journal of Statistical Planning and Inference · doi:10.1016/j.jspi.2009.01.003 · Cited by 74
“…a variety of penalized maximum likelihood or related estimators (e.g., Bunea (2004), Bunea & McKeague (2005), Fan & Li (2002, 2004), Li & Liang (2007), Wang & Leng (2007), Wang, G. Li and Jiang (2007), Wang, G. Li and Tsai (2007), Wang, R. Li and Tsai (2007), Yuan & Lin (2007), Zhang & Lu (2007)).”
Iterative Thresholding Algorithm for Sparse Inverse Covariance EstimationPreprint
Guillot, D., Rajaratnam, B., Rolfs, B. T., Maleki, A., & Wong, I. (2012) · arXiv (Cornell University) · doi:10.48550/arxiv.1211.2532 · Cited by 74
“Yuan and Lin [30] use interior point methods based on the max-det problem studied in [26].”
MethodInfluentialDifferential network analysis via lasso penalized D-trace loss
Yuan, H., Xi, R., Chen, C., & Deng, M. (2017) · Biometrika · doi:10.1093/biomet/asx049 · Cited by 87
Transfer Learning in Large-Scale Gaussian Graphical Models with False Discovery Rate Control
Li, S., Cai, T., & Li, H. (2022) · Journal of the American Statistical Association · doi:10.1080/01621459.2022.2044333 · Cited by 72
The Internet of Federated Things (IoFT)
Kontar, R., Shi, N., Yue, X., Chung, S., Byon, E., Chowdhury, M., Jin, J., Kontar, W., Masoud, N., Nouiehed, M., et al. (2021) · IEEE Access · doi:10.1109/access.2021.3127448 · Cited by 66
“the graphical modeling framework) is a problem that has received significant attention and research in the statistics and machine learning literature [167, 247, 319].”
BackgroundA sparse ising model with covariates
Cheng, J., Levina, E., Wang, P., & Zhu, J. (2014) · Biometrics · doi:10.1111/biom.12202 · Cited by 85
“…covariance matrix have appeared in recent years, with a focus on the highdimensional framework, for example, Meinshausen and Bühlmann (2006), Yuan and Lin (2007), d’Aspremont, Banerjee, and El Ghaoui (2008), Ravikumar et al. (2008), Rocha, Zhao, and Yu et al. (2008), Rothman et al. (2008),…”
BackgroundStatistical Significance of Clustering Using Soft Thresholding
Huang, H., Liu, Y., Yuan, M., & Marron, J. S. (2014) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2014.948179 · Cited by 96
“…sparse methods have been introduced to improve the estimation of the high dimensional covariance matrix, see e.g. Meinshausen and B¨uhlmann (2006); Yuan and Lin (2007); Friedman et al. (2008); Rothman et al. (2008); Fan et al. (2009); Witten et al. (2009); Yuan (2010); Cai et al. (2011); Danaher…”
MethodSelf-citationA multiple testing approach to the regularisation of large sample correlation matrices
Bailey, N. T. J., Pesaran, M. H., & Smith, L. V. (2018) · Journal of Econometrics · doi:10.1016/j.jeconom.2018.10.006 · Cited by 81
Identifying small mean-reverting portfolios
d’Aspremont, A. (2010) · Quantitative Finance · doi:10.1080/14697688.2010.481634 · Cited by 71
“A recent stream of works on the asymptotic consistency of these procedures can be found in Meinshausen & Yu (2007), Candes & Tao (2007), Banerjee et al. (2007), Yuan & Lin (2007) or Rothman, Bickel, Levina & Zhu (2007) among others.”
BackgroundSparse Matrix Inversion with Scaled LassoPreprint
Sun, T., & Zhang, C. (2012) · arXiv (Cornell University) · doi:10.48550/arxiv.1202.2723 · Cited by 72
“Lasso-type methods, or convex minimization algorithms with the l1 penalty on all entries of Θ∗, have been developed in [3, 14], and in [30] with l1 penalization on the off-diagonal matrix only.”
MethodSTORE: Sparse Tensor Response Regression and Neuroimaging AnalysisPreprint
Sun, W. W., & Li, L. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1609.04523 · Cited by 108
“Then one may employ sparse graphical model estimation (Yuan and Lin, 2007; Friedman et al., 2008; Zhang and Zou, 2014) and replace this inverse with a sparse precision matrix estimator.”
MethodStructural Pursuit Over Multiple Undirected Graphs
Zhu, Y., Shen, X., & Pan, W. (2014) · Journal of the American Statistical Association · doi:10.1080/01621459.2014.921182 · Cited by 75
“Methods to exploit matrix sparsity include [1, 5, 11, 13, 14, 15, 24], among others.”
BackgroundInference in High-Dimensional Graphical Models
Janková, J., & van de Geer, S. (2018) · CRC Press eBooks · doi:10.1201/9780429463976-14 · Cited by 87
“The graphical Lasso (see Yuan and Lin (2007), d’Aspremont, Banerjee, and El Ghaoui (2008), Friedman et al. (2008)) is based on the Gaussian log-likelihood function but regularizes it via an (cid:96) 1 -norm penalty on the off-diagonal elements of the precision matrix.”
BackgroundLearning Scale Free Networks by Reweighted L1 regularization
Liu, Q., & Ihler, A. (2011) · International Conference on Artificial Intelligence and Statistics · Cited by 88
Gaussian and bootstrap approximations for high-dimensional U-statistics and their applications
Chen, X. (2018) · The Annals of Statistics · doi:10.1214/17-aos1563 · Cited by 93
“In particular, motivation of this paper comes from the estimation and inference problems for large covariance matrix and its related functionals [55, 74, 64, 59, 73, 13, 10, 20, 21].”
BackgroundInferring sparse Gaussian graphical models with latent structure
Ambroise, C., Chiquet, J., & Matias, C. (2009) · Electronic Journal of Statistics · doi:10.1214/08-ejs314 · Cited by 85
“However, the approach for solving the problem in Yuan and Lin (2007) is not suited to high-dimensional settings, in contrast to the approach proposed in Banerjee et al. (2008).”
BackgroundMethodResultGraph estimation with joint additive models
Voorman, A., Shojaie, A., & Witten, D. (2013) · Biometrika · doi:10.1093/biomet/ast053 · Cited by 88
“As an alternative to (3), many authors have considered estimating Θ under the multivariate normality assumption by maximizing an `1-penalized joint log likelihood (see e.g. Yuan & Lin, 2007; Banerjee et al., 2008; Friedman et al., 2008).”
BackgroundMethodInfluentialJoint estimation of multiple high-dimensional precision matrices
Cai, T., Li, H., Liu, W., & Xie, J. (2015) · Statistica Sinica · doi:10.5705/ss.2014.256 · Cited by 85
“For the WS model, first a ring lattice of p vertex is created; one vertex is connected with its neighbors within order distance of 15, and then the edges of the lattice are rewired uniformly and randomly with a probability 0.01. These graph models have several topological properties such as sparcity and a “small world…”
BackgroundMethodInfluentialThe cluster graphical lasso for improved estimation of Gaussian graphical models
Tan, K. M., Witten, D., & Shojaie, A. (2014) · Computational Statistics & Data Analysis · doi:10.1016/j.csda.2014.11.015 · Cited by 86
“To overcome these problems, Yuan & Lin (2007) proposed to maximize the penalized log likelihood log det Θ − tr( SΘ ) − λ with respect to Θ , penalizing only the off-diagonal elements of Θ .”
BackgroundTranselliptical Graphical Models
Liu, H., Han, F., & Zhang, C. (2012) · neural information processing systems · Cited by 76
A Nonparametric Graphical Model for Functional Data With Application to Brain Networks Based on fMRI
Li, B., & Solea, E. (2017) · Journal of the American Statistical Association · doi:10.1080/01621459.2017.1356726 · Cited by 85
“Essentially, Gaussian graphical models rely on mutual linear regressions among random elements on the vertices—this is implicit in Yuan and Lin (2007) but explicit in some other Gaussian-assumption-based approaches such as Meinshausen and Bühlmann (2006), Friedman, Hastie, and Tibshirani (2008),…”
MethodSimultaneous Variable and Covariance Selection With the Multivariate Spike-and-Slab LASSO
Deshpande, S. K., Ročková, V., & George, E. I. (2019) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2019.1593179 · Cited by 82
“A particularly popular approach to estimating Ω is the graphical LASSO (GLASSO), which adds an 1 penalty to the negative log-likelihood of Ω (see, e.g., Yuan and Lin (2007), Banerjee et al. (2008), and Friedman et al. (2008)).”
High‐dimensional covariance matrix estimation
Lam, C. (2019) · Wiley Interdisciplinary Reviews Computational Statistics · doi:10.1002/wics.1485 · Cited by 90
Regularized Parameter Estimation in High-Dimensional Gaussian Mixture Models
Ruan, L., Yuan, M., & Zou, H. (2011) · Neural Computation · doi:10.1162/neco_a_00128 · Cited by 64
“We thus adopt the idea of a penalized likelihood estimate from Yuan and Lin (2007) and apply an 1-type penalty on the off-diagonal entries of the inverse covariance matrices.”
BackgroundMethodInfluentialSelf-citationLASSO-driven inference in time and space
Chernozhukov, V., Härdle, W. K., Huang, C., & Wang, N. (2021) · The Annals of Statistics · doi:10.1214/20-aos2019 · Cited by 78
Honest confidence regions and optimality in high-dimensional precision matrix estimation
Janková, J., & van de Geer, S. (2016) · Test · doi:10.1007/s11749-016-0503-5 · Cited by 84
“This approach is typically referred to as the graphical Lasso, and has been studied in detail in several papers, see e.g. Friedman et al. (2008), Rothman et al. (2008), Ravikumar et al. (2008) and Yuan and Lin (2007).”
MethodMetabolic connectivity for differential diagnosis of dementing disorders
Titov, D., Diehl‐Schmid, J., Shi, K., Perneczky, R., Zou, N., Grimmer, T., Li, J., Drzezga, A. E., & Yakushev, I. Y. (2016) · Journal of Cerebral Blood Flow & Metabolism · doi:10.1177/0271678x15622465 · Cited by 67
“SICE is of particular advantage for studies with small n.(24,25) However, due to the effect of shrinkage SICE is insufficiently robust for estimating values of non-zero entries, providing only a good estimation of (sparse) binary structure of ?̂”
BackgroundIdentifying gene regulatory network rewiring using latent differential graphical models
Tian, D., Gu, Q., & Ma, J. (2016) · Nucleic Acids Research · doi:10.1093/nar/gkw581 · Cited by 64
Bayesian Regularization for Graphical Models With Unequal Shrinkage
Gan, L., Narisetty, N. N., & Liang, F. (2018) · Journal of the American Statistical Association · doi:10.1080/01621459.2018.1482755 · Cited by 83
“(Yuan and Lin 2007; Baner-jee, El Ghaoui, and d’Aspremont 2008; Friedman, Hastie, and Tibshirani 2008; Fan, Feng, and Wu 2009).”
BackgroundMethodAdipocyte OGT governs diet-induced hyperphagia and obesity
Li, M., Vera, N. B., Yang, Y., Zhang, B., Ni, W., Ziso-Qejvanaj, E., Ding, S., Zhang, K., Yin, R., Wang, S., et al. (2018) · Nature Communications · doi:10.1038/s41467-018-07461-x · Cited by 68
A Unified Framework for Structured Graph Learning via Spectral ConstraintsPreprint
Kumar, S., Ying, J., Cardoso, J. V. D. M., & Palomar, D. P. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1904.09792 · Cited by 63
“…h(Θ) = ∑ i,j |Θij |, which promotes element-wise sparsity on the graph matrix Θ, is a common choice of regularization function to learn a sparse structure (Yuan and Lin, 2007; Shojaie and Michailidis, 2010a,b; Ravikumar et al., 2010; Mazumder and Hastie, 2012; Fattahi and Sojoudi, 2019).”
BackgroundInformation-theoretic bounds on model selection for Gaussian Markov random fields
Wang, W., Wainwright, M. J., & Ramchandran, K. (2010) · IEEE International Symposium on Information Theory (ISIT) · doi:10.1109/isit.2010.5513573 · Cited by 61
“[3], [4], [5], [6], [7]) have been shown to yield consistent estimators for Gaussian graphical models, or the associated inverse covariance matrices.”
BackgroundTao, M., Wang, Y., & Zhou, H. H. (2013) · The Annals of Statistics · doi:10.1214/13-aos1128 · Cited by 61
“Yuan and Lin (2007) considered large covariance matrix estimation in a Gaussian graph model.”
BackgroundDifferential network analysis: A statistical perspective
Shojaie, A. (2020) · Wiley Interdisciplinary Reviews Computational Statistics · doi:10.1002/wics.1508 · Cited by 77
“In this approach, first considered by (Yuan and Lin, 2007; Banerjee et al., 2008) and popularized by the efficient graphical lasso algorithm (Friedman et al.”
MethodA Path Following Algorithm for Sparse Pseudo-Likelihood Inverse Covariance Estimation (SPLICE)
Rocha, G. V., Zhao, P., & Yu, B. (2008) · Defense Technical Information Center (DTIC) · doi:10.21236/ada487557 · Cited by 69
“2.2 Comparison of exact and pseudo-likelihoods in the Gaussian case Banerjee et al. (2005), Yuan and Lin (2007), Banerjee et al. (2007) and Friedman et al. (2008) have considered estimates defined as the minimizers of the exact likelihood penalized by the (cid:96) 1 -norm of the off-diagonal terms of…”
BackgroundMethodInfluentialSun, L., Patel, R., Liu, J., Chen, K., Wu, T., Li, J., Reiman, E. M., & Ye, J. (2009) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/1557019.1557162 · Cited by 66
“Unlike most of the existing algorithms [2, 6, 12, 15, 17, 18], the proposed SICE algorithm estimates the inverse covariance matrix directly.”
MethodBayesian Modelling of Dupuytren Disease by Using Gaussian Copula Graphical Models
Mohammadi, R., Abegaz, F., van den Heuvel, E. R., & Wit, E. C. (2016) · Journal of the Royal Statistical Society Series C (Applied Statistics) · doi:10.1111/rssc.12171 · Cited by 73
Monitoring the covariance matrix via penalized likelihood estimation
Li, B., Wang, K., & Yeh, A. B. (2012) · IIE Transactions · doi:10.1080/0740817x.2012.663952 · Cited by 59
“Yuan and Lin (2007) proposed to add a Least Absolute Shrinkage and Selection Operator (LASSO)-based L1 penalty to the likelihood function when estimating the inverse of the covariance matrix in Gaussian graphical models.”
BackgroundPosterior convergence rates for estimating large precision matrices using graphical models
Banerjee, S., & Ghosal, S. (2014) · Electronic Journal of Statistics · doi:10.1214/14-ejs945 · Cited by 77
Innovated interaction screening for high-dimensional nonlinear classification
Fan, Y., Kong, Y., Li, D., & Zheng, Z. (2015) · The Annals of Statistics · doi:10.1214/14-aos1308 · Cited by 59
“See, for example, [1, 5, 12, 22, 28, 29, 31], among others.”
Weighted Lasso with Data Integration
Bergersen, L. C., Glad, I. K., & Lyng, H. (2011) · Statistical Applications in Genetics and Molecular Biology · doi:10.2202/1544-6115.1703 · Cited by 69
Robust estimation of precision matrices under cellwise contamination
Tarr, G., Müller, S., & Weber, N. C. (2015) · Computational Statistics & Data Analysis · doi:10.1016/j.csda.2015.02.005 · Cited by 58
“Yuan and Lin (2007) consider minimising the penalised negative log-likelihood, tr(SΘ) − log |Θ| + λ∑i, j |θi j|, (4) over the set of PD matrices where λ is a tuning parameter to control the amount of shrinkage.”
BackgroundMethodGraphical Models in Meshed Distribution Grids: Topology Estimation, Change Detection & Limitations
Deka, D., Talukdar, S., Chertkov, M., & Salapaka, M. V. (2020) · IEEE Transactions on Smart Grid · doi:10.1109/tsg.2020.2978541 · Cited by 69
“J vv ( i , j ) + J θθ ( i , j ) 0 if and only if ( ij ) is a true edge in G .”
Metadata
- Authors
- Yuan, M., & Lin, Y.
- Affiliations
- Georgia Institute of Technology (US)
- University of Wisconsin–Madison (US)
- Venue
- Biometrika
- Publisher
- Oxford University Press (OUP)
- Published
- 2007-02-07
- Volume
- 94
- Issue
- 1
- Pages
- 19-35
- ISSN
- 0006-3444 (ISSN-L)
- 1464-3510
- OpenAlex
- W2081746825
- Semantic Scholar
- 20ee888ca9c3bae784f266405b4d6395c6f32955
- Cited by
- 1,627
- References
- 16
- Citations by year
- 2026: 43 · 2025: 66 · 2024: 60 · 2023: 64 · 2022: 86 · 2021: 131 · 2020: 112 · 2019: 147 · 2018: 142 · 2017: 91 · 2016: 123 · 2015: 98 · 2014: 109 · 2013: 86 · 2012: 85
- FWCI
- 38.99
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- Biometrika (publishedVersion)
- RePEc: Research Papers in Economics (submittedVersion)
- HighWire Press Open Archive (submittedVersion)
- Funders
- National Science Foundation
- Topics
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods and Inference
- Physical Sciences › Mathematics › Statistics and Probability › Random Matrices and Applications
- Physical Sciences › Computer Science › Artificial Intelligence › Bayesian Modeling and Causal Inference
- Keywords
- Gaussian graphical models, model selection, penalized likelihood, sparse estimation, shrinkage estimation, hyperparameter optimization, covariance structure
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.