Retrieval and adaptation in CBR through Bayesian Network for diagnosis of hepatic pathologies

摘要: This article describes a modeling of knowledge for Case Based Reasoning system (CBR) applied to the diagnosis hepatic pathologies, where cases and domain are expressed by Bayesian network (BN). In fact, we interested in retrieval adaptation phases. The phase consists selecting most similar case log linear model considering Network as log-linear based on simplification probabilities. means modifying solutions retrieved fit current problem. dependence between these two phases is defined measures: similarity measure an measure. objective this guarantee case, which easiest way adapt improve performance CBR. An example pathologies will illustrate proposed approach.