A Belief Revision Framework For Probabilistic Object-Oriented Databases

Abstract

While past research extended the conventional object-oriented data model using probability theory to better capture the uncertainty of the real world, database modification has not been addressed to date.  This paper builds on previous work by Kornatzky and Shimony (1994) to address database modification for probabilistic object-oriented databases and proposes a framework for belief revision.  According to the framework, beliefs on data values are revised with the arrival of new information.  The proposed belief revision framework can be used to update data stored in an object-oriented database.