Learning Bayesian networks from data . Some applications in biomedicine

We will start the talk introducing Bayesian networks [15, 21] as one probabilistic graphical model that is based on the conditional independence concept between triplet of variables. This concept constitutes the semantic for Bayesian networks and is checked in the structure (a directed acyclic graph) by means of the u–separation criterion. The quantitative component of the paradigm (parameters) will be defined as conditional probabilities between the nodes (variables) in the directed acyclic graph. Taking into account both components, a factorization of the joint probability distribution is obtained. One advantage of this factorization is the reduction on the number of parameters needed in order to specify the joint distribution probability. However, the most important feature of this paradigm is that it can be used to infer in domains with inherent uncertainty. The reasoning inside the model consists in the propagation of the evidence through the model. This task was proved to be NP-hard in [4] for multiple connected networks. In the talk we will review one approximate algorithm to accomplish it, based on a simulation of the Bayesian network proposed in [13].

[1]  Nir Friedman,et al.  Bayesian Network Classifiers , 1997, Machine Learning.

[2]  Pedro Larrañaga,et al.  Feature Subset Selection by Bayesian network-based optimization , 2000, Artif. Intell..

[3]  Nir Friedman,et al.  Learning Belief Networks in the Presence of Missing Values and Hidden Variables , 1997, ICML.

[4]  Richard O. Duda,et al.  Pattern classification and scene analysis , 1974, A Wiley-Interscience publication.

[5]  Nir Friedman,et al.  Tissue classification with gene expression profiles , 2000, RECOMB '00.

[6]  Finn V. Jensen,et al.  Bayesian Networks and Decision Graphs , 2001, Statistics for Engineering and Information Science.

[7]  Michael J. Pazzani,et al.  Searching for Dependencies in Bayesian Classifiers , 1995, AISTATS.

[8]  J. Mesirov,et al.  Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. , 1999, Science.

[9]  David Heckerman,et al.  Knowledge Representation and Inference in Similarity Networks and Bayesian Multinets , 1996, Artif. Intell..

[10]  Pedro Larrañaga,et al.  Structure Learning of Bayesian Networks by Genetic Algorithms: A Performance Analysis of Control Parameters , 1996, IEEE Trans. Pattern Anal. Mach. Intell..

[11]  S. Lauritzen.  The EM algorithm for graphical association models with missing data , 1995 .

[12]  Judea Pearl,et al.  Probabilistic reasoning in intelligent systems , 1988 .

[13]  Pedro Larrañaga,et al.  Learning Recursive Bayesian Multinets for Data Clustering by Means of Constructive Induction , 2002, Machine Learning.

[14]  Pedro Larrañaga,et al.  Geographical clustering of cancer incidence by means of Bayesian networks and conditional Gaussian networks , 2001, AISTATS.

[15]  Max Henrion,et al.  Propagating uncertainty in bayesian networks by probabilistic logic sampling , 1986, UAI.

[16]  J. A. Lozano,et al.  Estimation of Distribution Algorithms: A New Tool for Evolutionary Computation , 2001 .

[17]  Pedro Larrañaga,et al.  Using Bayesian networks in the construction of a bi-level multi-classifier. A case study using intensive care unit patients data , 2001, Artif. Intell. Medicine.

[18]  Pedro Larrañaga,et al.  Learning Bayesian network structures by searching for the best ordering with genetic algorithms , 1996, IEEE Trans. Syst. Man Cybern. Part A.

[19]  Gregory F. Cooper,et al.  A Bayesian method for the induction of probabilistic networks from data , 1992, Machine Learning.