dRAP: A Framework for Distributed Mining Firts-Order Frequent Patterns

Warning

This publication doesn't include Faculty of Arts. It includes Faculty of Informatics. Official publication website can be found on muni.cz.

Authors	BLAŤÁK Jan POPELÍNSKÝ Lubomír
Year of publication	2006
Type	Article in Proceedings
Conference	16th International Conference on Inductive Logic Programming
MU Faculty or unit	Faculty of Informatics
Citation
Field	Informatics
Keywords	inductive logic programming; parallel and distributed mining
Description	In this paper dRAP, a framework for distributed mining first-order frequent patterns is presented. It extends the RAP system for running on parallel shared-nothing architecture. It utilizes several well-known methods for parallel mining propositional frequent patterns and new algorithm that minimizes communication overhead. We show that the new algorithm require significantly smaller number of messages passed than the other methods.
Related projects:	Dynamic Geovisualisation in Crises Management