190 likes | 357 Views
An Efficient Candidate Pruning Technique for High Utility Pattern Mining. Chowdhury Farhan Ahmed, Syed Khairuzzaman Tanbeer, Byeong-Soo Jeong, Young-Koo Lee PAKDD 2009. Outline. Motivation Definition Method Experimental Result Conclusion. Motivation.
E N D
An Efficient Candidate Pruning Technique for High Utility Pattern Mining Chowdhury Farhan Ahmed, Syed Khairuzzaman Tanbeer, Byeong-Soo Jeong, Young-Koo Lee PAKDD 2009
Outline Motivation Definition Method Experimental Result Conclusion
Motivation In this paper, we propose a novel tree-based candidate pruning technique HUC-Prune (high utility candidates prune) to efficiently mine high utility patterns.
Definition I:I = {i1, i2, ......,im} be a set of items D:transaction database T:transaction,T={T1,T2,…..,Tn}, Ti ∈ D is a subset of I.
HUC-prune minutil=106.75 twu(a)=44+37+74+102=257 twu(b)=347 twu(c)=89<minutil twu(d)=308 twu(e)=204
HUC-prune 38+102
HUC-prune All candidate:{b,e:140},{d,e:166},{e:204},{a,b:257}, {a,d:176},{a,b,d:176},{a:257},{b,d:228},{d:308},{b:347} High utility 6 itemsets:{a,b:134},{a,b,d:146},{b:132}, {b,d:154},{d:128},{d,e:134}
Conclusion Extensive performance analyses show that our technique is very efficient in highutility pattern mining and it outperforms the existing algorithms in both denseand sparse datasets.