BAYESIAN NETWORK REFINEMENT VIA MACHINE LEARNING APPROACH

A new approach to refining Bayesian network structures from new data i s developed. Most previous work has only considered the refinement of the network's conditional probability parameters and has not addressed the issue of refining the network's structure. We tackle this problem by a machine learning approach based on a formalism known as the Mini mum Description Length (MDL) principle. The MDL principle is well suit ed to this task since it can perform tradeoffs between the accuracy. s implicity, and closeness to the existent structure. Another salient fe ature of this refinement approach is the capability of refining a netw ork structure using partially specified data. Moreover, a localization scheme is developed for efficient computation of the description leng ths since direct evaluation involves exponential time resources.