英语论文网

egments.Some limitations of the suggested methodology are asfollows. With regard to the number of target data setswhich should be compared, the methodology is suitable foronly two. If there are three or more data sets to becompared over time, then a recursive methodology willhave to be developed. The methodology is run on data setswhich have discrete values. If there is a data set which hascontinuous values, then a preprocessing step to makediscrete values is needed.As a further research area, we plan to extend ourmethodology to discover changes over three or more datasets. It will be a promising research area to set up acampaign management plan based on our suggestedmethodology, and it will also be interesting to check theeffectiveness of the campaign.AcknowledgementHee Seok Song contributed as a corresponding author forthis paper.ReferencesAGRAWAL, R. and G. PSAILA (1995) Active data mining, inProceedings of the First International Conference on KnowledgeDiscovery and Data Mining (KDD-95), 3–8.BAY, S.D. and M.J. PAZZANI (1999) Detecting change in categoricaldata: mining contrast sets, in Proceedings of the Fifth InternationalConference on Knowledge Discovery and Data Mining(KDD-99), 302–306.CHEUNG, D.W., J. HAN, V.T. NG and C.Y. WONG (1996)Maintenance of discovered association rules in large databases:an incremental updating technique, in Proceedings of the TwelfthInternational Conference on Data Engineering (ICDE-96),106–114.DAS, G., D.GUNOPULOUS and H.MANNILA (1997) Finding similartime series, in Proceedings of the First European Symposium:Principles of DataMining and Knowledge Discovery (PKDD ’97),88–100.DAS, G., K. LIN, H.MANNILA, G.RENGANATHAN and P. SMYTH(1998) Rule discovery from time series, in Proceedings of theFourth International Conference on Knowledge Discovery andData Mining (KDD-98), 16–22.DONG, G. and J. LI (1999) Efficient mining of emerging patterns:discovering trends and differences, in Proceedings of the FifthInternational Conference on Knowledge Discovery and DataMining (KDD-99), 43–52.FELDMAN, R., Y. AUMANN, A. AMIR and H. MANILA (1997)Efficient algorithms for discovering frequent sets in incrementaldatabases, in SIGMOD ’97 Workshop on Research Issues in DataMining and Knowledge Discovery (DMKD ’97), 59–66.FREUND, Y. and Y. MANSOUR (1997) Learning under persistentdrift, in Computational Learning Theory: Third EuropeanConference, 109–118.GANTI, V., J. GEHRKE and R. RAMAKRISHNAN (1999) A frameworkfor measuring changes in data characteristics, in Proceedingsof the Eighteenth ACM SIGACT-SIGMOD-SIGARTSymposium on Principles of Database Systems (PODS-99),126–137.HAN, J. and M. KAMBER (2001) Data Mining: Concepts andTechniques, San Francisco, CA: Morgan Kaufmann, 200–207.HAN, J., G. DONG and Y. YIN (1999) Efficient mining of partialperiodic patterns in time series database, in Proceedings of theFifteenth International Conference on Data Engineering (ICDE1999), 106–115.HELMBOLD, D.P. and P.M. LONG (1994) Tracking drifting conceptsby minimizing disagreements, Machine Learning, 14, 27–45.HUSSAIN, F., H. LIU, E. SUZUKI and H. LU (2000) Exception rulemining with a relative interestingness measure, in Proceedings ofPacific Asia Conference on Knowledge Discovery in Databases(PAKDD), 86–97.LANQUILLON, C. (1999) Information filtering in changing domains,in Proceedings of the International Joint Conference on ArtificialIntelligence (IJCAI99), 41–48.LI, J., G. DONG and K. RAMAMOHANARAO (2000) Making use ofthe most expressive jumpi