Další formáty:
BibTeX
LaTeX
RIS
@article{562245, author = {Blaťák, Jan and Popelínský, Lubomír}, article_location = {Praha}, article_number = {5}, keywords = {knowledge discovery in databases; inductive logic programming; frequent patterns; feature construction; propositionalization}, language = {eng}, issn = {1210-0552}, journal = {Neural Network World}, title = {Mining first-order maximal frequent patterns}, volume = {14}, year = {2004} }
TY - JOUR ID - 562245 AU - Blaťák, Jan - Popelínský, Lubomír PY - 2004 TI - Mining first-order maximal frequent patterns JF - Neural Network World VL - 14 IS - 5 SP - 381-390 EP - 381-390 PB - UIVT AV ČR SN - 12100552 KW - knowledge discovery in databases KW - inductive logic programming KW - frequent patterns KW - feature construction KW - propositionalization N2 - Frequent patterns discovery is one of the most important data mining tasks. We introduce RAP, the first system for finding first-order maximal frequent patterns. We describe search strategies and methods of pruning the search space. RAP generates long patterns much faster than other systems.RAP has been used for feature construction for propositional as well as multirelational data. We prove that partial search for maximal frequent patterns as new features is competitive with other approaches and results in classification accuracy increase. ER -
BLAŤÁK, Jan a Lubomír POPELÍNSKÝ. Mining first-order maximal frequent patterns. \textit{Neural Network World}. Praha: UIVT AV ČR, 2004, roč.~14, č.~5, s.~381-390. ISSN~1210-0552.
|