Reversible MCMC on Markov equivalence classes of sparse directed acyclic graphs

@article{He2012ReversibleMO,
  title={Reversible MCMC on Markov equivalence classes of sparse directed acyclic graphs},
  author={Yangbo He and Jinzhu Jia and Bin Yu},
  journal={ArXiv},
  year={2012},
  volume={abs/1209.5860}
}
Author(s): He, Y; Jia, J; Yu, B | Abstract: Graphical models are popular statistical tools which are used to represent dependent or causal complex systems. Statistically equivalent causal or directed graphical models are said to belong to a Markov equivalent class. It is of great interest to describe and understand the space of such classes. However, with currently known algorithms, sampling over such classes is only feasible for graphs with fewer than approximately 20 vertices. In this paper… 

Figures and Tables from this paper

Counting and exploring sizes of Markov equivalence classes of directed acyclic graphs

TLDR
The proposed size counting is efficient for Markov equivalence classes of sparse DAGs with hundreds of vertices and exploration of the size and edge distributions finds that the sizes of sparse classes grow approximately exponentially with the numbers of Vertices.

Learning Markov Equivalence Classes of Directed Acyclic Graphs: An Objective Bayes Approach

TLDR
This work approaches the problem of model selection among noncausal sparse Gaussian DAGs by directly scoring EGs, using an objective Bayes method, and proposes a Markov Chain Monte Carlo (MCMC) strategy to explore the space of EGs using sparsity constraints, and illustrates the performance of the method on simulation studies, as well as on a real dataset.

Bayesian Model Selection of Gaussian Directed Acyclic Graph Structures

TLDR
This paper reviews some recent results on Bayesian model selection of Gaussian DAG models under a unified framework and introduces a general MCMC scheme that can be adopted both for model selection and EGs together with a couple of applications on real data sets.

Counting and Sampling from Markov Equivalent DAGs Using Clique Trees

TLDR
This paper proposes a new technique for counting the number of DAGs in a Markov equivalence class based on the clique tree representation of chordal graphs and shows that in the case of bounded degree graphs, the proposed algorithm is polynomial time.

Learning Markov Equivalence Classes of Gaussian DAGs via Observational and Interventional Data: an Objective Bayes Approach

TLDR
This paper presents a finer partition of DAGs into equivalence classes, each one represented by an Interventional Essential Graph (I-EG), also called Completed Partially Directed Graph (CPDAG), and proposes several methods based on Markov chains to avoid such limit.

Partition MCMC for Inference on Acyclic Digraphs

TLDR
A novel algorithm is proposed, which employs the underlying combinatorial structure of DAGs to define a new grouping, and convergence is improved compared to structure MCMC, while still retaining the property of producing an unbiased sample.

Equivalence class selection of categorical graphical models

A Bayesian Approach to Directed Acyclic Graphs with a Candidate Graph

TLDR
A Bayesian method, baycn (BAYesian Causal Network), to estimate the posterior probability of three states for each edge, which achieves high accuracy with small variation across different scenarios and is comparable or better than existing Bayesian methods.

Approximate Counting of Graphical Models via MCMC Revisited

  • J. Peña
  • Computer Science, Mathematics
    CAEPIA
  • 2013
TLDR
The approximate ratio of essential graphs (EGs) to DAGs for up to 31 nodes is calculated, and it is suggested that Lauritzen–Wermuth–Frydenberg chain graphs are considerably more expressive than D AGs.

Complexity analysis of Bayesian learning of high-dimensional DAG models and their equivalence classes

TLDR
A rapid mixing result for a random walk Metropolis-Hastings algorithm is proved using a canonical path method, which reveals that the complexity of Bayesian learning of sparse equivalence classes grows only polynomially in n and p, under some common high-dimensional assumptions.

References

SHOWING 1-10 OF 42 REFERENCES

A characterization of Markov equivalence classes for acyclic digraphs

Undirected graphs and acyclic digraphs (ADGs), as well as their mutual extension to chain graphs, are widely used to describe dependencies among variables in multivariate distributions. In

The size distribution for Markov equivalence classes of acyclic digraph models

Bayesian model averaging and model selection for markov equivalence classes of acyclic digraphs

Acyclic digraphs (ADGs) are widely used to describe dependences among variables in multivariate distributions. In particular, the likelihood functions of ADG models admit convenient recursive

Formulas for counting acyclic digraph Markov equivalence classes

Approximate Counting of Graphical Models via MCMC Revisited

  • J. Peña
  • Computer Science, Mathematics
    CAEPIA
  • 2013
TLDR
The approximate ratio of essential graphs (EGs) to DAGs for up to 31 nodes is calculated, and it is suggested that Lauritzen–Wermuth–Frydenberg chain graphs are considerably more expressive than D AGs.

Equivalence and Synthesis of Causal Models

TLDR
The canonical representation presented here yields an efficient algorithm for determining when two embedded causal models reflect the same dependency information, which leads to a model theoretic definition of causation in terms of statistical dependencies.

Learning Equivalence Classes of Bayesian-Network Structures

TLDR
It is argued that it is often appropriate to search among equivalence classes of network structures as opposed to the more common approach of searching among individual Bayesian-network structures, and a convenient graphical representation for an equivalence class of structures is described and a set of operators that can be applied to that representation by a search algorithm to move among equivalENCE classes are introduced.

Learning Essential Graph Markov Models From Data

TLDR
In a model selection procedure where many models are to be compared, computational efficiency is critical, and a scoring metric can be developed for EGs with categorical (multinomial) data.

Optimal Structure Identification With Greedy Search

TLDR
This paper proves the so-called "Meek Conjecture", which shows that if a DAG H is an independence map of another DAG G, then there exists a finite sequence of edge additions and covered edge reversals in G such that H remains anindependence map of G and after all modifications G =H.

Robust graphical modeling of gene networks using classical and alternative t-distributions

TLDR
It is demonstrated that penalized likelihood inference combined with an application of the EM algorithm provides a computationally efficient approach to model selection in the $t-distribution case.