Database Mining: Propositionalization for Knowledge Discovery in Relational Databases

Propositionalization is a process that leads from relational data and background knowledge to a single-table representation thereof, which serves as the input to widespread systems for knowledge discovery in databases. Systems for propositionalization thus support the analyst during the usually costly phase of data preparation for data mining. We present a framework that allows for a unified description of approaches to propositionalization. Within our framework, we systematically enhance existing approaches with techniques well-known in the area of relational databases, especially the application of aggregate functions. Further, we suggest special database schema transformations for efficiency reasons. We put special emphasis on empirical investigations into the spectrum of approaches. Data set sizes vary across different orders of magnitude, up to several million data points. Also, the domains are diverse, ranging from biological data sets to financial ones, in order to demonstrate the broad applicability of propositionalization.

Mark-André Krogel, Dr.-Ing, Dipl.-Inf., M.Sc.: studied Applied Computer Science (Otto-von-Guericke Universität Magdeburg) and Cognitive Science (The University of Edinburgh), won KDD Cup 2001 (San Francisco) with Propositionalization, now Test Engineer for interfaces to electronic financial markets at icubic AG (Magdeburg, Frankfurt/Main)