Skip to main content

Induction of first-order rules and constraints for knowledge discovery

Adaptivity and the ability to make use of knowledge implicitly present in low-level data are of primary importance for enhancing business and administrative processes in Europe. With data warehousing in relational databases becoming more and more popular, there is increasing awareness that structural or relational learning as investigated in the field of Inductive Logic Programming (ILP) is a key technique for knowledge discovery. The main goals of this project were to enhance existing techniques and develop new ILP algorithms that can be used as efficient tools for solving hard problems, using induced first-order classification rules and integrity constraints.

Staff and Students

Peter Flach.



Nada Lavrac, Blaz Zupan and Ljupco Todorovski. (Jozef Stefan Institute, Slovenia).


This research was supported by the Royal Society under a Joint Project with Central and Eastern Europe.