摘要: The aim of relational learning is to develop methods for the induction descriptions in representation formalisms that are more expressive than attribute-value representation. Feature terms have been studied formalize object-centered declarative languages and can be seen as a subset first-order logic. We present formalism based on feature we show how performed natural way using notion subsumption an informational ordering. Moreover also allow specify incomplete information way. An example such inductive methods, indie, presented. indie performs bottom-up heuristic search lattice term space. Results this method several domains explained.