Using New Data to Re ne a Bayesian Network

We explore the issue of re ning an exis tent Bayesian network structure using new data which might mention only a subset of the variables Most previous works have only considered the re nement of the net work s conditional probability parameters and have not addressed the issue of re n ing the network s structure We develop a new approach for re ning the network s structure Our approach is based on the Minimal Description Length MDL princi ple and it employs an adapted version of a Bayesian network learning algorithm de veloped in our previous work One of the adaptations required is to modify the previ ous algorithm to account for the structure of the existent network The learning algo rithm generates a partial network structure which can then be used to improve the exis tent network We also present experimental evidence demonstrating the e ectiveness of our approach