Inductive logic programming (ILP) provides tools for relational data mining, that is, mining from data stored in multiple tables. It works with the powerful language of logic programmes, both as prior domain knowledge and as describing the discovered patterns. A good introduction to the theory, implementation, and applications of ILP is written by Muggleton and De Raedt (1994). Another introduction to ILP that also relates logic programming terminology to database terminology, is given by Dzeroski and Lavrac (2001). Books that address ILP have been written by Lloyd (2003); De Raedt (2005,1996), and Furukawa et al. (1999).
The basic data mining task of ILP is as follows: Given positive (and possibly negative) examples, a concept description language, and possibly background knowledge, find a set of association rules that covers most of the positive examples but only few of the negative examples.