Mutual Information Preconditioning Improves Bayesian Networks Learning of Medical Databases

Bayesian Networks represent one of the most successful tools for medical diagnosis and therapies follow-up. We present an algorithm for Bayesian network structure learning, that is a variation of the standard search-and-score approach. The proposed algorithm overcomes the creation of redundant network structures that may include non significant connections between variables. In particular, the algorithm finds what relationships between the variables must be prevented, by exploiting the binarization of a square matrix containing the mutual information (MI) among all pairs of variables. Two different binarization methods are implemented. The first one is based on the MRMR (maximum relevance minimum redundancy) selection strategy. The second one uses a threshold. The MI binary matrix is exploited as a pre-conditioning step for the subsequent greedy search procedure that optimizes the network score, reducing the number of possible search paths in the greedy search. Ouralgorithm has been tested on two different medical datasets and compared against the standard search-and-score algorithm as implemented in the DEAL package.