Inference Bayesian Network for Multi-topographic neural network communication: a case study in documentary data
Résumé
In this paper we present an original approach consisting in assimilating the behavior of the MultiSOM model, whose core model represents a significant extension of the classical Kohonen SOM model, to the one model of a Bayesian inference network. This approach is used both for validating the MultiSOM inter-map communication principles and for enhancing the accuracy of the probabilistic correlation computation mode that is already provided by the model In a complementary way, our approach also led us to prove that a neural multi-map model provided with unsupervised learning might well behave as a Bayesian inference network in which the estimation of posterior probabilities becomes a simple process only using prior similarity measures.