ECAI 2004 Conference Paper

[full paper]

Weak OI: Ideal Refinement of Datalog Clauses using Primary Keys

Siegfried Nijssen, Joost N. Kok

Inductive Logic Programming (ILP) algorithms are frequently used for data mining tasks in multi-relational databases. However, by most ILP algorithms primary key information is disregarded while this information is often available for such databases. This work shows how primary key information can be incorporated in a downward refinement operator. We show how primary keys can be used to define sublanguages of full clausal logic and provide evidence that one can define ideal refinement operators for these languages. We incorporate our refinement operator in a multirelational data mining algorithm to demonstrate its feasibility.

Keywords: Data Mining, Knowledge Representation, Inductive Logic Programming, Machine Learning

Citation: Siegfried Nijssen, Joost N. Kok: Weak OI: Ideal Refinement of Datalog Clauses using Primary Keys. In R.López de Mántaras and L.Saitta (eds.): ECAI2004, Proceedings of the 16th European Conference on Artificial Intelligence, IOS Press, Amsterdam, 2004, pp.520-524.