Blankets joint posterior score for learning Markov network structures
From MaRDI portal
Publication:1687296
Abstract: Markov networks are extensively used to model complex sequential, spatial, and relational interactions in a wide range of fields. By learning the structure of independences of a domain, more accurate joint probability distributions can be obtained for inference tasks or, more directly, for interpreting the most significant relations among the variables. Recently, several researchers have investigated techniques for automatically learning the structure from data by obtaining the probabilistic maximum-a-posteriori structure given the available data. However, all the approximations proposed decompose the posterior of the whole structure into local sub-problems, by assuming that the posteriors of the Markov blankets of all the variables are mutually independent. In this work, we propose a scoring function for relaxing such assumption. The Blankets Joint Posterior score computes the joint posterior of structures as a joint distribution of the collection of its Markov blankets. Essentially, the whole posterior is obtained by computing the posterior of the blanket of each variable as a conditional distribution that takes into account information from other blankets in the network. We show in our experimental results that the proposed approximation can improve the sample complexity of state-of-the-art scores when learning complex networks, where the independence assumption between blanket variables is clearly incorrect.
Recommendations
- Improving structure MCMC for Bayesian networks through Markov blanket resampling
- Using Markov blankets for causal structure learning
- Marginal pseudo-likelihood learning of discrete Markov network structures
- Structure learning of contextual Markov networks using marginal pseudo-likelihood
- The IBMAP approach for Markov network structure learning
- Approximate structure learning for large Bayesian networks
- scientific article; zbMATH DE number 5183798
Cites work
- scientific article; zbMATH DE number 1646249 (Why is no real title available?)
- scientific article; zbMATH DE number 1702339 (Why is no real title available?)
- scientific article; zbMATH DE number 1817585 (Why is no real title available?)
- scientific article; zbMATH DE number 3423418 (Why is no real title available?)
- scientific article; zbMATH DE number 3787827 (Why is no real title available?)
- scientific article; zbMATH DE number 107482 (Why is no real title available?)
- scientific article; zbMATH DE number 1409213 (Why is no real title available?)
- Causation, prediction, and search
- Decomposition of maximum likelihood in mixed graphical interaction models
- Efficient Markov Network Structure Discovery Using Independence Tests
- Hyper Markov laws in the statistical analysis of decomposable graphical models
- Improving Markov network structure learning using decision trees
- Local causal and Markov blanket induction for causal discovery and feature selection for classification. Part I: Algorithms and empirical evaluation
- Machine learning in complex networks
- Markov Network-Based Unified Classifier for Face Recognition
- Probabilistic graphical models.
- Some Methods for Strengthening the Common χ 2 Tests
- The IBMAP approach for Markov network structure learning
- The Libra toolkit for probabilistic models
- The University of Florida sparse matrix collection
Describes a project that uses
Uses Software
This page was built for publication: Blankets joint posterior score for learning Markov network structures
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1687296)
