A Report on the Summer School on
Relational Data Mining
17-18 August 2002, Helsinki, Finland

Saso Dzeroski and Bernard Zenko

Relational Data Mining (RDM) is the multi-disciplinary field dealing with knowledge discovery from relational databases consisting of multiple tables. To emphasize the contrast to typical data mining approaches that look for patterns in a single relation of a database, the name Multi-Relational Data Mining (MRDM) is often used as well. Mining data which consists of complex/structured objects also falls within the scope of this field: the normalized representation of such objects in a relational database requires multiple tables. The field aims at integrating results from existing fields such as inductive logic programming (ILP), KDD, data mining, machine learning and relational databases; producing new techniques for mining multi- relational data; and practical applications of such techniques. Present RDM approaches consider all of the main data mining tasks, including association analysis, classification, clustering, learning probabilistic models and regression. The pattern languages used by single-table data mining approaches for these data mining tasks have been extended to the multiple-table case. Relational pattern languages now include relational association rules, relational classification rules, relational decision trees, and probabilistic relational models, among others. RDM algorithms have been developed to mine for patterns expressed in relational pattern languages. Typically, data mining algorithms have been upgraded from the single-table case: for example, distance-based algorithms for prediction and clustering have been upgraded by defining distance measures between examples/instances represented in relational logic. RDM methods have been successfully applied across many application areas, ranging from the analysis of business data, through bioinformatics (including the analysis of complete genomes) and pharmacology (drug design) to Web mining (e.g., information extraction from Web sources). The Summer School on Relational Data Mining provided a comprehensive introduction to the techniques and applications of relational data mining by leading experts in the field. The lectures given at the school are summarized below. The slides of the lectures were published as handouts and are also available for download at the Web page of the school http://www-ai.ijs.si/SasoDzeroski/RDMSchool . The Summer School concluded with a panel discussion, where the lecturers and participants raised a number of interesting issues concerning the future of relational data mining. Interesting statements from the panel include: "The future of RDM is in upgrading (probabilistic models, SVMs, neural networks) and downgrading (to make it more efficient for specific applications, e.g., on sequences) (Luc De Raedt)"; "The different approaches to RDM should be integrated" (Hendrik Blockeel); "RDM should be more of an engineering discipline than art" (Thomas Gaertner); "The European and US scientific communities working on RDM topics should communicate and not develop in isolation" (Lise Getoor). The Summer School was organized by the Jozef Stefan Institute,
Ljubljana, with the help and support of the University of Helsinki. It was financially supported by ILPnet2 (The Network of Excellence in Inductive Logic Programming). It was attended by 34 participants from 13 countries (including the USA and New Zealand).

Created by LaTeX2html on 2002-11-04.