Bayesian Networks for Knowledge Discovery in Large Medical Data Set

This paper discusses the modeling using the combination of rough sets,rule based reasoning and Bayesian net- work(BN)based on the large amounts of data in hospital information system.On the basis of attributes reduction algo- rithm of rough set,the proposed method takes synthetically into account the influences of rule-based reasoning.The limi- tation of attribute variable in information tables was compressed.The minimal attributes was obtained via the compression of attribute columns.Due to the acquisition of minimal attributes,the complexity of BN structure was largdy decreased; probability reasoning could he realized by BN.The efficiency of this method is validated by the practical examples.