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 301–400 of 1,000
Detecting positive correlations in a multivariate sample
Ery Arias-castro, A., Bubeck, S., & Lugosi, G. (2015) · Bernoulli · doi:10.3150/13-bej565 · Cited by 38
“Cai et al. (2010); El Karoui (2008) and inference on sparse graphical models, for example Lam and Fan (2009); Meinshausen and Bu¨hlmann (2006); Rajaratnam et al. (2008); Verzelen and Villers (2010); Yuan and Lin (2007). While the literature has focused almost exclusively on estimating the dependency structure, we focu…”
BackgroundHigh-dimensional Sparse Inverse Covariance Estimation using Greedy MethodsPreprint
Johnson, C. C., Jalali, A., & Ravikumar, P. (2011) · arXiv (Cornell University) · doi:10.48550/arxiv.1112.6411 · Cited by 46
“In this section, we compare our global and local greedy methods to the `1-regularized Gaussian MLE, analyzed in [10], and to `1-regularization (Lasso) based neighborhood selection, analyzed in [8, 14].”
BackgroundMethodInfluentialPRISMA: PRoximal Iterative SMoothing AlgorithmPreprint
Orabona, F., Argyriou, A., & Srebro, N. (2012) · HAL · doi:10.48550/arxiv.1206.2372 · Cited by 37
GGM Knockoff Filter: False Discovery Rate Control for Gaussian Graphical Models
Li, J., & Maathuis, M. H. (2021) · Journal of the Royal Statistical Society Series B (Statistical Methodology) · doi:10.1111/rssb.12430 · Cited by 46
“e false discovery proportion (FDP). Previous research on structure learning of Gaussian graphical models has mainly focused on handling high-dimensional settings(e.g., Meinshausen and Buhlmann, 2006; Yuan and Lin, 2007;¨ Friedman et al., 2008; d’Aspremont et al., 2008; Rothman et al., 2008; Raskutti et al., 2009; Fan…”
BackgroundOn inferring presence of an individual in a mixture: a Bayesian approach
Clayton, D. (2010) · Biostatistics · doi:10.1093/biostatistics/kxq035 · Cited by 37
Sparse Nonparametric Graphical Models
Lafferty, J., Liu, H., & Wasserman, L. (2012) · Statistical Science · doi:10.1214/12-sts391 · Cited by 58
“This estimator goes back at least to Yuan and Lin (2007), and an iterative lasso algorithm for doing the optimization was first proposed by Banerjee, El Ghaoui and d’Aspremont (2008). In our experiments we use the R packages glasso (Friedman, Hastie and Tibshirani, 2007) and huge to implement this algorithm.”
MethodA Scalable Multi-Resolution Spatio-Temporal Model for Brain Activation and Connectivity in Fmri Data
Castruccio, S., Ombao, H., & Genton, M. G. (2018) · Biometrics · doi:10.1111/biom.12844 · Cited by 38
“This penalized likelihood approach on the inverse covariance was proposed by Yuan and Lin (2007) as an application of the maxdet problem (Vandenberghe et al., 1998); a faster approach for high-dimensional covariance estimation, the graphical LASSO (Friedman et al., 2008), has become widely popular…”
MethodPenalized Covariance Matrix Estimation Using a Matrix-Logarithm Transformation
Deng, X., & Tsui, K. (2012) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2012.715556 · Cited by 38
“Research in this direction includes Meinshausen and Buhlmann (2006), Banerjee et al. (2006), Huang et al. (2006), Yuan and Lin (2007), Friedman et al. (2008), Rajaratnam et al. (2008), Levina et al. (2008), Peng et al. (2009), Cai et al. (2010) , Yuan (2010), and Zhou et al. (2010) among many…”
BackgroundMethodBayesian composite quantile regression
Huang, H., & Chen, Z. (2015) · Journal of Statistical Computation and Simulation · doi:10.1080/00949655.2015.1014372 · Cited by 33
Identifying Gene Network Rewiring by Integrating Gene Expression and Gene Network Data
Xu, T., Ou-Yang, L., Hu, X. T., & Zhang, X.-F. (2018) · IEEE Transactions on Computational Biology and Bioinformatics · doi:10.1109/tcbb.2018.2809603 · Cited by 36
“Gaussian Graphical Models (GGMs) are widely used to model gene networks [3], [4], [5].”
MethodScale calibration for high-dimensional robust regression
Loh, P. (2021) · Electronic Journal of Statistics · doi:10.1214/21-ejs1936 · Cited by 40
Change-Point Detection for Graphical Models in the Presence of Missing Values
Londschien, M., Kovács, S., & Bühlmann, P. (2020) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2020.1853549 · Cited by 45
“Computational approaches for the estimator were also presented by Banerjee et al. (2008), theoretical properties of it were investigated by Yuan and Lin (2007), Rothman et al. (2008), Lam and Fan (2009) and Ravikumar et al. (2011).”
BackgroundMethodPenalized estimation in high-dimensional hidden Markov models with state-specific graphical models
Städler, N., & Mukherjee, S. (2013) · The Annals of Applied Statistics · doi:10.1214/13-aoas662 · Cited by 39
“;X) denotes the observed log-likelihood and pen( K,λ) is a penalty function involving the 1-norms of the inverse covariance matrices [Friedman, Hastie and Tibshirani (2008), Meinshausen and Bühlmann (2006), Yuan and Lin (2007)] that we describe in detail below.”
BackgroundMethodLearning a common substructure of multiple graphical Gaussian models
Hara, S., & Washio, T. (2012) · Neural Networks · doi:10.1016/j.neunet.2012.11.004 · Cited by 38
“…algorithms for this do not scale to high dimensional data, the scope of studies has shifted to a relaxed setting (Meinshausen and B¨uhlmann, 2006; Yuan and Lin, 2007; Banerjee et al., 2008), where Covariance Selection is formulated as a convex optimization problem using a ℓ 1 -regularization…”
MethodInfluentialXie, B., Pan, W., & Shen, X. (2009) · Bioinformatics · doi:10.1093/bioinformatics/btp707 · Cited by 33
“In addition, there will be some computational issues in implementing the constraint that the resulting covariance matrix estimate is positive definite (Huang et al., 2006; Yuan and Lin, 2007).”
Equivalence of graphical lasso and thresholding for sparse graphs
Sojoudi, S. (2014) · Allerton Conference on Communication, Control, and Computing · doi:10.1109/allerton.2014.7028566 · Cited by 33
“A simple analysis of this equation leads to the conditions provided in (3).”
BackgroundMethodInfluentialLi, Z., Lee, K., Karagas, M. R., Madan, J. C., Hoen, A. G., O’Malley, A. J., & Li, H. (2018) · Statistics in Biosciences · doi:10.1007/s12561-018-9219-2 · Cited by 35
Joint estimation of multiple dependent Gaussian graphical models with applications to mouse genomics
Xie, Y., Liu, Y., & Valdar, W. S. J. (2016) · Biometrika · doi:10.1093/biomet/asw035 · Cited by 36
“More recently, numerous papers have proposed estimation using sparse penalized maximum likelihood (Yuan & Lin, 2007; Banerjee et al., 2008; d’Aspremont et al., 2008; Rothman et al., 2008; Ravikumar et al., 2011).”
BackgroundNetwork Modeling in Biology: Statistical Methods for Gene and Brain Networks
Wang, Y. X. R., Li, L., Li, J. J., & Huang, H. (2020) · Statistical Science · doi:10.1214/20-sts792 · Cited by 38
“A rich wealth of literature exists both in the context of gene expression analysis [119, 142, 160] and general high dimensional inference [248, 62, 256, 81] to tackle this problem.”
BackgroundBrownlees, C. T., Nualart, E., & Sun, Y. (2018) · Journal of Applied Econometrics · doi:10.1002/jae.2642 · Cited by 41
Reconstruction of sparse connectivity in neural networks from spike train covariances
Pernice, V., & Rotter, S. (2013) · Journal of Statistical Mechanics Theory and Experiment · doi:10.1088/1742-5468/2013/03/p03008 · Cited by 32
“We use the entry-wise L 1 matrix norm of the off-diagonal elements of ˆ B to define a cost function Γ [29].”
MethodMulti-Domain Sampling With Applications to Structural Inference of Bayesian Networks
Zhou, Q. (2011) · Journal of the American Statistical Association · doi:10.1198/jasa.2011.ap10346 · Cited by 39
“Other authors propose to solve the optimization problem using an adaptation of classical optimization methods, such as interior point (Yuan and Lin 2007) or block coordinate descent methods (Banerjee et al. 2008; Friedman et al. 2007).”
BackgroundMethodGaussian and robust Kronecker product covariance estimation: Existence and uniqueness
Soloveychik, I., & Трушин, Д. В. (2016) · Journal of Multivariate Analysis · doi:10.1016/j.jmva.2016.04.001 · Cited by 32
“Non-linear structures are also quite common in engineering applications, among them are the Kronecker product model [24–27], linear inverse covariance structures such as graphical models, [28, 29] and others.”
BackgroundMultitask Quantile Regression Under the Transnormal Model
Fan, J., Xue, L., & Zou, H. (2015) · Journal of the American Statistical Association · doi:10.1080/01621459.2015.1113973 · Cited by 32
Das, A., Sampson, A. L., Lainscsek, C. S. M., Muller, L., Lin, W., Doyle, J. C., Cash, S. S., Halgren, E., & B, T. J. S. A. (2017) · Neural Computation · doi:10.1162/neco_a_00936 · Cited by 38
“…Hastie, & Tibshirani, 2008; Hsieh, Sustik, Dhillon, & Ravikumar, 2011; Hsieh, Sustik, Dhillon, Ravikumar, & Poldrack, 2013; Oztoprak, Nocedal, Rennie, & Olsen, 2012; Rothman, Bickel, Levina, & Zhu, 2008; Scheinberg, Ma, & Goldfarb, 2010; Yuan & Lin, 2007) have been proposed for the q = 1 case.”
A general algorithm for covariance modeling of discrete data
Popovic, G. C., Hui, F. K. C., & Warton, D. I. (2017) · Journal of Multivariate Analysis · doi:10.1016/j.jmva.2017.12.002 · Cited by 36
“Gaussian graphical models [2, 13, 28, 36, 41] for example, describe conditional independence relationships between variables, which can be used to distinguish between direct and indirect relationships among variables.”
BackgroundPetralia, F., Wang, L., Peng, J., Yan, A., Zhu, J., & Wang, P. (2018) · Bioinformatics · doi:10.1093/bioinformatics/bty280 · Cited by 30
“There are many recent works on applying high-dimensional GGMs under sparsity and/or structural assumptions to infer gene coexpression networks (Friedman et al., 2008; Meinshausen and Bühlmann, 2006; Schäfer and Strimmer, 2005; Yuan and Lin, 2007), including our previous works (Cheng et al.”
Shrinkage tuning parameter selection in precision matrices estimation
Lian, H. (2011) · Journal of Statistical Planning and Inference · doi:10.1016/j.jspi.2011.03.008 · Cited by 36
“To compare the performance of different selec-tors, we use the Frobenius norm || Ω 0 − ˆ Ω F as well as the entropy loss (Yuan and Lin, 2007) defined by For the performance in terms of sparsity of the matrix, we use false positives (number of zero entries in the true precision matrix identified to be…”
MethodGeneralized Stability Approach for Regularized Graphical ModelsPreprint
Müller, C. L., Bonneau, R., & Kurtz, Z. D. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1605.07072 · Cited by 34
“Gaussian graphical models have become popular after the advent of computationally tractable estimators, such as neighborhood selection (Meinshausen and Bühlmann, 2006) and sparse inverse covariance estimation (Banerjee et al., 2008; Yuan and Lin, 2007).”
BackgroundGene Regulation Network Inference With Joint Sparse Gaussian Graphical Models
Chun, H., Zhang, X., & Zhao, H. (2014) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2014.956876 · Cited by 32
“The sparse GGM estimation has been extensively studied recently including Meinshausen and Buhlmann (2006); Yuan and Lin (2007); Peng et al. (2009); Lam and Fan (2009); and Guo et al. (2011).”
MethodEfficient Computation of ℓ1Regularized Estimates in Gaussian Graphical Models
Yuan, M. (2008) · Journal of Computational and Graphical Statistics · doi:10.1198/106186008x382692 · Cited by 39
“In particular, Yuan and Lin (2007a) showed that graphGarrote is consistent in selecting the graphical model if p is held fixed as n goes to infinity whereas graphLasso may not be.”
BackgroundMethodInfluentialSelf-citationVogels, L., Mohammadi, R., Schoonhoven, M., & Birbil, Ş. İ. (2024) · Journal of the American Statistical Association · doi:10.1080/01621459.2024.2395504 · Cited by 37
“In the case of (cid:7) 1 -norm regularization, the penalty term is given by p ( K ) = (cid:4) i , j , i (cid:2)= j | k ij | and the optimization problem (5) becomes a convex program that Yuan and Lin (2007) solve using the interior point algorithm (Vandenberghe, Boyd, and Wu 1998).”
High dimensional posterior convergence rates for decomposable graphical models
Xiang, R., Khare, K., & Ghosh, M. (2015) · Electronic Journal of Statistics · doi:10.1214/15-ejs1084 · Cited by 36
Dimensionality reduction in forecasting with temporal hierarchies
Nystrup, P., Lindström, E., Møller, J. K., & Madsen, H. (2021) · International Journal of Forecasting · doi:10.1016/j.ijforecast.2020.12.003 · Cited by 36
Joint Gaussian graphical model estimation: A survey
Tsai, K. H., Koyejo, S., & Kolar, M. (2022) · Wiley Interdisciplinary Reviews Computational Statistics · doi:10.1002/wics.1582 · Cited by 39
“Hence, a typical strategy to estimate Ω in a high-dimensional setting is to add a regularization function, such as the `1-norm of the parameters, to the log-likelihood function (1), which encourages the graph to be sparse or have other desirable structural biases (Yuan and Lin, 2007; Bühlmann and Van De Geer, 2011).”
BackgroundEstimating heterogeneous graphical models for discrete data with an application to roll call voting
Guo, J., Cheng, J., Levina, E., Michailidis, G. C., & Zhu, J. (2015) · The Annals of Applied Statistics · doi:10.1214/13-aoas700 · Cited by 36
“A number of methods have been recently proposed in the literature to fit sparse Gaussian graphical models [see, e.g., Banerjee, El Ghaoui and d’Aspremont (2008), Meinshausen and Bühlmann (2006), Peng, Zhou and Zhu (2009), Rothman et al. (2008), Yuan and Lin (2007), Ravikumar et al. (2011) and references therein].”
BackgroundMethodInfluentialEstimating sufficient reductions of the predictors in abundant high-dimensional regressions
Cook, R. D., Forzani, L., & Rothman, A. J. (2012) · The Annals of Statistics · doi:10.1214/11-aos962 · Cited by 30
“…λ . b The use of L 1 -penalized likelihood optimization to estimate a sparse inverse covariance matrix has been studied extensively in the literature [d’Aspremont, Banerjee and El Ghaoui (2008), Yuan and Lin (2007), Friedman, Hastie and Tibshirani (2008), Rothman et al. (2008), Lam and Fan (2009)].”
MethodInfluentialA general family of trimmed estimators for robust high-dimensional data analysis
Yang, E., Lozano, A. C., & Aravkin, A. Y. (2018) · Electronic Journal of Statistics · doi:10.1214/18-ejs1470 · Cited by 32
“The most widely used estimator, the Graphical Lasso minimizes the negative Gaussian log-likelihood regularized by the 1 norm of the entries (or the off-diagonal entries) of the precision matrix (see [52, 17, 3]).”
MethodEstimating the error variance in a high-dimensional linear model
Yu, G., & Bien, J. (2019) · Biometrika · doi:10.1093/biomet/asz017 · Cited by 30
Robust Gaussian Graphical Modeling Vial1Penalization
Sun, H., & Li, H. (2012) · Biometrics · doi:10.1111/j.1541-0420.2012.01785.x · Cited by 37
“…fact that the number of genes is usually larger than the sample size, regularization methods have been developed in recent years to estimate high dimensional Gaussian graphical models (Meinshausen and B¨uhlmann, 2006; Yuan and Lin, 2007; Friedman, Hastie, and Tibshirani, 2008; Peng et al., 2009).”
BackgroundInference in High Dimensions with the Penalized Score TestPreprint
Voorman, A., Shojaie, A., & Witten, D. (2014) · arXiv (Cornell University) · doi:10.48550/arxiv.1401.2678 · Cited by 30
“Further, the framework for (cid:96) 1 -penalized M -estimators mentioned in Section 5.2, in principle, applies to the graphical lasso (Yuan and Lin, 2007b; Friedman et al., 2008) and (cid:96) 1 -penalized generalized linear models (Friedman et al., 2010).”
MethodInfluentialLearning Gene Regulatory Networks from Next Generation Sequencing Data
Jia, B., Xu, S., Xiao, G., Lamba, V., & Liang, F. (2017) · Biometrics · doi:10.1111/biom.12682 · Cited by 34
“For gLasso and nodewise regression, the simulated ZINB data first went through the logarithm transformation and nonparanormal transformation, which have been widely used in RNA-seq data analysis, and then the methods were applied.”
MethodInfluentialMultiple Matrix Gaussian Graphs Estimation
Zhu, Y., & Li, L. (2018) · Journal of the Royal Statistical Society Series B (Statistical Methodology) · doi:10.1111/rssb.12278 · Cited by 33
“A large number of statistical methods have been proposed to estimate a sparse Gaussian graphical model (Meinshausen and B¨uhlmann (2006), Yuan and Lin (2007), Friedman et al. (2008), Ravikumar et al. (2011) and Cai et al. (2011), among others).”
Recent Advances on Penalized Regression Models for Biological Data
Wang, P., Chen, S., & Yang, S. (2022) · Mathematics · doi:10.3390/math10193695 · Cited by 31
Calibrated Precision Matrix Estimation for High-Dimensional Elliptical Distributions
Zhao, T., & Liu, H. (2014) · IEEE Transactions on Information Theory · doi:10.1109/tit.2014.2360980 · Cited by 29
“Banerjee et al. (2008); Yuan and Lin (2007); Friedman et al. (2008) propose the penalized Gaussian log-likelihood method named graphical lasso (GLASSO), Tuo Zhao and Han Liu are affiliated with Department of Operations Research Research and Financial Engineering, Princeton University, Princeton, NJ…”
MethodBerry-Esseen bounds for estimating undirected graphs
Wasserman, L., Kolar, M., & Rinaldo, A. (2014) · Electronic Journal of Statistics · doi:10.1214/14-ejs928 · Cited by 32
“There are many methods for estimating undirected graphs, such as the glasso (Yuan and Lin, 2007; Friedman and Tibshirani, 2007) and sparse parallel regression (Meinshausen and B¨uhlmann, 2006).”
MethodInferring metabolic networks using the Bayesian adaptive graphical lasso with informative priors
Peterson, C. B., Vannucci, M., Karakas, C., Choi, W. T.-S., Ma, L., & Maletić‐Savatić, M. (2013) · Statistics and Its Interface · doi:10.4310/sii.2013.v6.n4.a12 · Cited by 32
“For example, the graphical lasso achieves sparsity in the estimation of the precision matrix Ω by imposing a penalty on its L1 norm [17, 34, 51].”
Safo, S. E., Li, S., & Long, Q. (2017) · Biometrics · doi:10.1111/biom.12715 · Cited by 32
“atrices are identity matrices, but we relax this assumption as it can be overly restrictive in practice. In particular, our method allows the use of sparse covariance matrices (Friedman et al., 2007; Yuan and Lin, 2007) from which the underlying structural network may be inferred. In section 2, we present the proposed…”
MethodAIC for the Lasso in generalized linear models
Ninomiya, Y., & Kawano, S. (2016) · Electronic Journal of Statistics · doi:10.1214/16-ejs1179 · Cited by 29
An improved modified cholesky decomposition approach for precision matrix estimation
Kang, X., & Deng, X. (2019) · Journal of Statistical Computation and Simulation · doi:10.1080/00949655.2019.1687701 · Cited by 34
“Specifically, for a given hard thresholding δ l , l = 1 , . . . , H , the corresponding BIC( δ l ) [2] is computed by where ˜ Ω δ l = (˜ ω 0, and ˜ e ij ( l ) = 1 otherwise.”
MethodInfluentialLe, K. T., Chaux, C., Richard, F. J. P., & Guedj, É. (2020) · Computational Statistics & Data Analysis · doi:10.1016/j.csda.2020.107031 · Cited by 29
Goodness-of-fit tests for high-dimensional Gaussian linear models
Verzélen, N., & Villers, F. (2010) · The Annals of Statistics · doi:10.1214/08-aos629 · Cited by 36
“[11] and Yuan and Lin [27] extend this method to infer directly the inverse covariance ma trix by minimizing the log-likelihood penalized by the ll norm.”
MethodGroup Sparse Priors for Covariance EstimationPreprint
Marlin, B. M., Schmidt, M., & Murphy, K. P. (2012) · arXiv (Cornell University) · doi:10.48550/arxiv.1205.2626 · Cited by 29
“One particularly promising regularization approach to covariance estimation is to penalize the ℓ 1 -norm of the precision matrix, Ω = Σ − 1 , to encourage sparsity in the precision matrix (Banerjee et al., 2008; Friedman et al., 2007; Yuan and Lin, 2007; Duchi et al., 2008; Schmidt et al., 2009).”
BackgroundMethodRobust Multimodal Graph Matching: Sparse Coding Meets Graph MatchingPreprint
Fiori, M., Sprechmann, P., Vogelstein, J. T., Musé, P., & Sapiro, G. (2013) · arXiv (Cornell University) · doi:10.48550/arxiv.1311.6425 · Cited by 31
“In numerous applications, this matrix is known to be sparse, and in this regard the graphical Lasso has proven to be a good estimator for the inverse covariance matrix (Yuan & Lin, 2007; Fiori et al., 2012) (also for non-Gaussian data (Loh & Wainwright, 2012)).”
BackgroundConfidence regions for high-dimensional generalized linear models under sparsityPreprint
Janková, J., & van de Geer, S. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1610.01353 · Cited by 29
“iterature for the case of quadratic loss; this is the same problem as estimation of the edge weights in undirected graphical models (see e.g. Friedman et al. (2008), Meinshausen and Buhlmann¨ (2006), Yuan and Lin (2007)). The challenge arises because, in general, Θ depends on β 0 (an exception is the quadratic loss).…”
MethodNetwork Reconstruction via the Minimum Description Length Principle
Peixoto, T. P. (2025) · Physical Review X · doi:10.1103/physrevx.15.011065 · Cited by 38
Structural Graphical Lasso for Learning Mouse Brain Connectivity
Yang, S., Sun, Q., Ji, S., Wonka, P., Davidson, I. A. K., & Ye, J. (2015) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/2783258.2783391 · Cited by 28
“cision matrix), since the nonzero off-diagonal elements of the precision matrix represent edges in the graph [4, 35].”
BackgroundGrafting-light: fast, incremental feature selection and structure learning of Markov random fields
Zhu, J., Lao, N., & Xing, E. P. (2010) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/1835804.1835845 · Cited by 36
“Markov random fields (GMRFs), inverse covariance estimation methods [2, 32 , 3] have been developed based on the � 1-norm penalized maximum likelihood estimation.”
MethodEstimation of large covariance and precision matrices from temporally dependent observations
Shu, H., & Nan, B. (2019) · The Annals of Statistics · doi:10.1214/18-aos1716 · Cited by 39
“Various algorithms for the `1 penalized maximum likelihood method (`1-MLE) and its variants have been developed [5, 35, 44, 75], and related theoretical properties have been investigated by [48, 59, 60].”
MethodJournal of Machine Learning Research
Zhu, H., Strawn, N., & Dunson, D. B. (2016) · VTechWorks (Virginia Tech) · Cited by 28
Wang, C., & Jiang, B. (2019) · Computational Statistics & Data Analysis · doi:10.1016/j.csda.2019.106812 · Cited by 32
“One of the well-known and popular method in precision matrix estimation is referred to as the graphical lasso (Yuan and Lin, 2007; Banerjee et al., 2008; Friedman et al., 2008).”
MethodEfficient Hamiltonian, structure and trace distance learning of Gaussian states
Fanizza, M., Rouzé, C., & França, D. S. (2026) · Nature Communications · doi:10.1038/s41467-026-77405-3 · Cited by 30
“A variety of ef fi cient and practical algorithms is available 30,59 – 65 having computation complexity polynomial in the number of variables.”
Tu, J.-J., Ou-Yang, L., Zhu, Y., Yan, H., Qin, H., & Zhang, X.-F. (2021) · Bioinformatics · doi:10.1093/bioinformatics/btab502 · Cited by 32
Zhou, L., Wang, L., & Ogunbona, P. (2014) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2014.396 · Cited by 30
“This method is termed sparse inverse covariance estimation (SICE), or called graphical LASSO [5, 13] in pattern recognition.”
MethodInfluentialBayesian sparse graphical models for classification with application to protein expression data
Baladandayuthapani, V., Talluri, R., Ji, Y., Coombes, K. R., Lu, Y., Hennessy, B. T., Davies, M. A., & Mallick, B. K. (2014) · The Annals of Applied Statistics · doi:10.1214/14-aoas722 · Cited by 29
“…over the past few years in both frequentist [Bickel and Levina (2008), Chaudhuri, Drton and Richardson (2007), Friedman, Hastie and Tibshirani (2008), Meinshausen and Bühlmann (2006), Yuan and Lin (2007)] and Bayesian settings [Carvalho and Scott (2009), Giudici and Green (1999), Roverato (2002)].”
BackgroundJoint conditional Gaussian graphical models with multiple sources of genomic data
Chun, H., Chen, M., Li, B., & Zhao, H. (2013) · Frontiers in Genetics · doi:10.3389/fgene.2013.00294 · Cited by 27
“Finding a sparse precision matrix with various regularizations such as lasso and adaptive lasso (Tibshirani, 1996; Zou, 2006) has been studied by many researchers including Li and Gui (2006); Yuan and Lin (2007); Friedman et al. (2008).”
BackgroundZhao, T., Roeder, K., & Liu, H. (2013) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2013.858633 · Cited by 32
“There are two major approaches for estimating high-dimensional Gaussian graphical models: (i) graphical lasso (Banerjee, Ghaoui, and d’Aspremont 2008; Yuan and Lin 2007; Friedman, Hastie, and Tibshirani 2007), which maximizes the (cid:2) 1 -penalized Gaussian likelihood and simultaneously estimates…”
BackgroundMethodHigh dimensional Gaussian copula graphical model with FDR control
He, Y., Zhang, X., Wang, P., & Zhang, L. (2016) · Computational Statistics & Data Analysis · doi:10.1016/j.csda.2016.06.012 · Cited by 28
Modeling uncertainty in macroeconomic growth determinants using Gaussian graphical models
Dobra, A., Eicher, T. S., & Lenkoski, A. (2009) · Statistical Methodology · doi:10.1016/j.stamet.2009.11.003 · Cited by 31
“[44] Yuan, M.”
Villers, F., Schaeffer, B., Bertin, C., & Huet, S. (2008) · Statistical Applications in Genetics and Molecular Biology · doi:10.2202/1544-6115.1371 · Cited by 27
Robust sparse Gaussian graphical modeling
Hirose, K., Fujisawa, H., & Sese, J. (2017) · Journal of Multivariate Analysis · doi:10.1016/j.jmva.2017.07.012 · Cited by 36
“To estimate the sparse inverse covariance matrix, Yuan and Lin (2007) proposed minimizing the following penalized negative log-likelihood function: `λ(θ) = `(θ) + λ 2 ‖Ω− diag(Ω)‖1, where `(θ) is a negative log-likelihood function given by `(θ) = − ∑n i=1 log f(xi;θ), and λ ≥ 0 is a tuning parameter which controls the…”
BackgroundMethodInfluentialGraph-Valued RegressionPreprint
Liu, H., Chen, X., Lafferty, J., & Wasserman, L. (2010) · arXiv (Cornell University) · doi:10.48550/arxiv.1006.3972 · Cited by 27
Learning Local Dependence In Ordered DataPreprint
Yu, G., & Bien, J. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1604.07451 · Cited by 30
“This is a hierarchical group lasso penalty (Yuan & Lin 2007, Zhao et al. 2009, Jenatton et al. 2011) with group structure conveyed in Figure 1.”
BackgroundDAGs with No Curl: An Efficient DAG Structure Learning ApproachPreprint
Yu, Y., Gao, T., Yin, N., & Ji, Q. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.07197 · Cited by 106
“Learning DAG structures for continuous variables is often limited to Gaussian models (Yuan and Lin, 2007; Foygel and Drton, 2010; Schmidt et al., 2007; Mohan et al., 2012; Mohammadi et al., 2015).”
BackgroundQi, S., Wang, R., Zhang, T., Yang, X., Sun, R., & Wang, L. (2024) · IEEE/CAA Journal of Automatica Sinica · doi:10.1109/jas.2024.124341 · Cited by 31
Convergence and sparsity of Lasso and group Lasso in high-dimensional generalized linear models
Wang, L., You, Y., & Lian, H. (2014) · Statistical Papers · doi:10.1007/s00362-014-0609-3 · Cited by 26
Sundaram, P., Luessi, M., Bianciardi, M., Stufflebeam, S. M., Hämäläinen, M., & Solo, V. (2019) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2019.2962517 · Cited by 27
“We used the BIC criterion proposed by [28] to select λ . where S is the sample covariance matrix and − 1 λ is the inverse covariance estimator for a given λ .”
MethodConditional Independence Testing in Hilbert Spaces with Applications to Functional Data Analysis
Lundborg, A. R., Shah, R. D., & Peters, J. (2022) · Journal of the Royal Statistical Society Series B (Statistical Methodology) · doi:10.1111/rssb.12544 · Cited by 43
Matrix normal PCA for interpretable dimension reduction and graphical noise modeling
Zhang, C., Gai, K., & Zhang, S. (2024) · Pattern Recognition · doi:10.1016/j.patcog.2024.110591 · Cited by 26
On an Additive Semigraphoid Model for Statistical Networks With Application to Pathway Analysis
Li, B., Chun, H., & Zhao, H. (2014) · Journal of the American Statistical Association · doi:10.1080/01621459.2014.882842 · Cited by 26
“For recent advances in this area, see Meinshausen and Buhlmann (2006), Yuan and Lin (2007), Bickel and Levina (2008), Fried-man et al. (2008), Peng et al. (2009), Guo et al. (2009), Lam and Fan (2009).”
BackgroundMethodEstimation of variances and covariances for high‐dimensional data: a selective review
Tong, T., Wang, C., & Wang, Y. (2014) · Wiley Interdisciplinary Reviews Computational Statistics · doi:10.1002/wics.1308 · Cited by 29
“Other regularization methods include, for example, Yuan and Lin 91 , d’Aspremont et al. 26 , Friedman et al. 43 , Ravikumar et al. 66 , and references therein.”
Petralia, F., Song, W.-M., Tu, Z., & Wang, P. (2016) · Journal of Proteome Research · doi:10.1021/acs.jproteome.5b00925 · Cited by 25
Lü, S., Liu, Y., Yin, L., & Zhang, K. (2016) · Journal of the Royal Statistical Society Series B (Statistical Methodology) · doi:10.1111/rssb.12184 · Cited by 26
“Some common choices are the inverse of the sample covariance matrix and the estimate of the precision matrix computed by the banding method (Bickel and Levina, 2008) or the penalized likelihood method (Yuan and Lin, 2007; Friedman et al., 2008).”
MethodOn Sparse Nonparametric Conditional Covariance Selection
Kolar, M., Parikh, A. P., & Xing, E. P. (2018) · KiltHub Repository · doi:10.1184/r1/6476171 · Cited by 26
Sure Screening for Gaussian Graphical ModelsPreprint
Luo, S., Song, R., & Witten, D. (2014) · arXiv (Cornell University) · doi:10.48550/arxiv.1407.7819 · Cited by 26
“Perhaps the best-known among these is the graphical lasso (Friedman et al., 2008; Yuan and Lin, 2007), which is the solution to the optimization problem maximizeΘ{log det Θ− trace((XTX/n)Θ)− λ ∑ i 6=j |Θij |}.”
BackgroundRobust and sparse correlation matrix estimation for the analysis of high-dimensional genomics data
Serra, A., Coretto, P., Fratello, M., & Tagliaferri, R. (2018) · Bioinformatics · doi:10.1093/bioinformatics/btx642 · Cited by 25
Bayesian Regression With Network Prior: Optimal Bayesian Filtering Perspective
Qian, X., & Dougherty, E. R. (2016) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2016.2605072 · Cited by 26
“They typically enforce prior distribution assumptions on model coefficients β, including traditional multivariate Gaussian prior assumptions [20]–[23] as well as other sparse prior assumptions, including spike-and-slab prior, Bernoulli-Gaussian prior, Laplace prior, and other structured sparse prior assumptions arisin…”
BackgroundLearning Spatial-Temporal Varying Graphs with Applications to Climate Data Analysis
Chen, X., Liu, Y., Liu, H., & Carbonell, J. G. (2010) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v24i1.7658 · Cited by 25
“…of the most promising techniques for structure learning, especially for data with inherent sparse graph structures (Meinshausen and Bühlmann 2006; Yuan and Lin 2007) and have been successfully applied in diverse areas, such as gene regulatory network discovery (Friedman 2004), social network…”
MethodInfluentialElementary Estimators for Graphical Models
Yang, E., Lozano, A. C., & Ravikumar, P. (2014) · neural information processing systems · Cited by 31
Learning directed acyclic graphical structures with genetical genomics data
Gao, B., & Cui, Y. (2015) · Bioinformatics · doi:10.1093/bioinformatics/btv513 · Cited by 27
Grechkin, M., Fazel, M., Witten, D., & Lee, S. (2015) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v29i1.9636 · Cited by 25
“In order to obtain a sparse estimate for Σ−1, a number of authors have considered the graphical lasso optimization problem (Yuan and Lin 2007; Banerjee, El Ghaoui, and d’Aspremont 2008; Friedman, Hastie, and Tibshirani 2007):”
BackgroundLafit, G., Tuerlinckx, F., Myin‐Germeys, I., & Ceulemans, E. (2019) · Scientific Reports · doi:10.1038/s41598-019-53795-x · Cited by 25
Estimation of Gaussian graphs by model selection
Giraud, C. (2008) · Electronic Journal of Statistics · doi:10.1214/08-ejs228 · Cited by 25
“Huang et al. [13] or Yuan and Lin [21] (see also Banerjee et al. [1] and Friedman et al. [11]) suggest in turn to rather estimate C − 1 by minimizing the log-likelihood for the concentration matrix penalized by the l 1 -norm.”
BackgroundOn Semiparametric Exponential Family Graphical ModelsPreprint
Yang, Z., Ning, Y., & Liu, H. (2014) · arXiv (Cornell University) · doi:10.48550/arxiv.1412.8697 · Cited by 25
“To estimate Θ in high dimensions, Yuan and Lin (2007); Banerjee et al. (2008); Friedman et al. (2008) propose the penalized log-likelihood method named graphical LASSO.”
MethodEstimation of Gaussian graphs by model selectionPreprint
Giraud, C. (2007) · HAL · Cited by 25
A Graphical Model of Smoking-Induced Global Instability in Lung Cancer
Wang, Y., Qian, W., & Yuan, B. (2016) · IEEE Transactions on Computational Biology and Bioinformatics · doi:10.1109/tcbb.2016.2599867 · Cited by 24
“, possible biological interactions or regulations) among individual genes [15],[53].”
BackgroundCovariance estimation via sparse Kronecker structures
Leng, C., & Pan, G. (2018) · Bernoulli · doi:10.3150/17-bej980 · Cited by 24
“We fit the penalized Gaussian graphical model in Yuan and Lin [23] for estimating a sparse −1 and the sparse correlation matrix estimation method in Cui, Leng and Sun [8] for estimating a sparse .”
MethodStructural similarity and difference testing on multiple sparse Gaussian graphical models
Liu, W. (2017) · The Annals of Statistics · doi:10.1214/17-aos1539 · Cited by 31
“There is a rich and growing literature on learning GGMs in various settings; see, for example, Meinshausen and Buhlmann (2006), Yuan and Lin (2007), Friedman, Hastie and Tibshirani (2008), d’Aspremont, Banerjee and El Ghaoui (2008), Rothman et al. (2008), Fan, Feng and Wu (2009), Ravikumar et al.…”
BackgroundMethodNarayan, M., & Allen, G. I. (2016) · Frontiers in Neuroscience · doi:10.3389/fnins.2016.00108 · Cited by 28
“…such as sparse regularized regression (Meinshausen and Bühlmann, 2006), sparse penalized maximum likelihood (ML) or the graphical lasso (Yuan and Lin, 2007; Friedman et al., 2008) and others (Cai et al., 2011; Zhou et al., 2011) can be used to estimate ( i ) in our subject level model (Equation…”
MethodLatent Network Estimation and Variable Selection for Compositional Data Via Variational EM
Osborne, N., Peterson, C. B., & Vannucci, M. (2021) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2021.1935971 · Cited by 24
“Frequentist approaches often use a penalized likelihood to obtain sparse estimates of the regression coefficients, and include methods such as LASSO (Tibshirani 1996), adaptive LASSO (Zou 2006), and SCAD (Fan and Li 2001).”
MethodInfluential
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.