Individuals, relations and structures in probabilistic models

Relational data is equivalent to non-relational struc-tured data. It is this equivalence which permitsprobabilistic models of relational data. Learningof probabilistic models for relational data is possi-ble because one item of structured data is generallyequivalent to many related data items. Successionand inclusion are two relations that have been wellexplored in the statistical literature. A descriptionof the relevant statistical approaches is given. Therepresentation of relational data via Bayesian netsis examined, and compared with PRMs. The pa-per ends with some cursory remarks on structuredobjects