Discovery of frequent DATALOG patterns

  • Authors:
  • Luc Dehaspe;Hannu Toivonen

  • Affiliations:
  • Department of Computer Science, Katholieke Universiteit Leuven, Celestijnenlaan 200A, B-3001 Heverlee, Belgium. luc.dehaspe@cs.kuleuven.ac.be;Rolf Nevanlinna Institute & Department of Computer Science, P.O. Box 4, FIN-00014 University of Helsinki, Finland. hannu.toivonen@rni.helsinki.fi

  • Venue:
  • Data Mining and Knowledge Discovery
  • Year:
  • 1999

Quantified Score

Hi-index 0.00

Visualization

Abstract

Discovery of frequent patterns has been studied in a variety of datamining settings. In its simplest form, known from association rule mining,the task is to discover all frequent itemsets, i.e., all combinations of itemsthat are found in a sufficient number of examples. The fundamental taskof association rule and frequent set discovery has been extended invarious directions, allowing more useful patterns to be discovered withspecial purpose algorithms. We present WARMR, a general purpose inductive logicprogramming algorithm that addresses frequent querydiscovery: a very general DATALOG formulation of the frequent pattern discovery problem.The motivation for this novel approach is twofold. First, exploratorydata mining is well supported: WARMRoffers the flexibility required to experimentwith standard and in particular novel settings not supported by special purpose algorithms. Also, application prototypes basedon WARMR can be used as benchmarks in the comparison and evaluation ofnew special purpose algorithms. Second, the unified representation givesinsight to the blurred picture of the frequentpattern discovery domain. Within the DATALOG formulation a number ofdimensions appear that relink diverged settings.We demonstrate the frequent query approach and its use on twoapplications, one in alarm analysis, and one in a chemical toxicologydomain.