Mixed Graphical Models via Exponential Families

Markov Random Fields, or undirected graphical models are widely used to model highdimensional multivariate data. Classical instances of these models, such as Gaussian Graphical and Ising Models, as well as recent extensions (Yang et al., 2012) to graphical models specified by univariate exponential families, assume all variables arise from the same distribution. Complex data from high-throughput genomics and social networking for example, often contain discrete, count, and continuous variables measured on the same set of samples. To model such heterogeneous data, we develop a novel class of mixed graphical models by specifying that each node-conditional distribution is a member of a possibly different univariate exponential family. We study several instances of our model, and propose scalable M -estimators for recovering the underlying network structure. Simulations as well as an application to learning mixed genomic networks from next generation sequencing and mutation data demonstrate the versatility of our methods.

[1]  Seunghak Lee,et al.  Adaptive Multi-Task Lasso: with Application to eQTL Detection , 2010, NIPS.

[2]  Larry A. Wasserman,et al.  Stability Approach to Regularization Selection (StARS) for High Dimensional Graphical Models , 2010, NIPS.

[3]  Michael I. Jordan.  Graphical Models , 1998 .

[4]  L. Stein,et al.  A human functional protein interaction network and its application to cancer data analysis , 2010, Genome Biology.

[5]  A. Dobra,et al.  Copula Gaussian graphical models and their application to modeling functional disability data , 2011, 1108.1680.

[6]  N. Meinshausen,et al.  High-dimensional graphs and variable selection with the Lasso , 2006, math/0608017.

[7]  Genevera I. Allen,et al.  A Local Poisson Graphical Model for Inferring Networks From Sequencing Data , 2013, IEEE Transactions on NanoBioscience.

[8]  Larry A. Wasserman,et al.  High Dimensional Semiparametric Gaussian Copula Graphical Models. , 2012, ICML 2012.

[9]  Steven J. M. Jones,et al.  Comprehensive molecular portraits of human breast tumours , 2013 .

[10]  J. Besag.  Spatial Interaction and the Statistical Analysis of Lattice Systems , 1974 .

[11]  Desmond J. Smith,et al.  A genome-wide map of human genetic interactions inferred from radiation hybrid genotypes. , 2010, Genome research.

[12]  C. O. A. D. P. R. M. A. E. Stimation.  Covariate Adjusted Precision Matrix Estimation with an Application in Genetical Genomics , 2011 .

[13]  M. Frydenberg,et al.  Decomposition of maximum likelihood in mixed graphical interaction models , 1989 .

[14]  Tianxi Li,et al.  High-Dimensional Mixed Graphical Models , 2013, 1304.2810.

[15]  S. Lauritzen.  Propagation of Probabilities, Means, and Variances in Mixed Graphical Association Models , 1992 .

[16]  J. Lafferty,et al.  High-dimensional Ising model selection using ℓ1-regularized logistic regression , 2010, 1010.0311.

[17]  N. Wermuth,et al.  Graphical Models for Associations between Variables, some of which are Qualitative and some Quantitative , 1989 .

[18]  Pradeep Ravikumar,et al.  Graphical Models via Generalized Linear Models , 2012, NIPS.

[19]  Trevor J. Hastie,et al.  Learning Mixed Graphical Models , 2012, ArXiv.

[20]  Andrew McCallum,et al.  Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data , 2001, ICML.

[21]  H. Zou,et al.  Regularized rank-based estimation of high-dimensional nonparanormal graphical models , 2012, 1302.3082.

[22]  Pradeep Ravikumar,et al.  Conditional Random Fields via Univariate Exponential Families , 2013, NIPS.

[23]  Pradeep Ravikumar,et al.  On Poisson Graphical Models , 2013, NIPS.

[24]  S. Lauritzen,et al.  Mixed graphical association models; discussions and reply , 1989 .