Title :
N-Most Interesting Closed Itemset Mining
Author :
Songram, Panida ; Boonjing, Veera
Author_Institution :
Dept. of Math. & Comput. Sci., King Mongkut´´s Inst. of Technol., Bangkok
Abstract :
In this paper, we propose an efficient algorithm, NCLOSED, for mining the N k-closed itemsets with the highest supports for 1 up to a certain kmax value. The algorithm adopts best-first search strategy to generate closed itemsets with highest remaining supports. It does not keep closed itemsets mined in main memory to ensure that they are really closed. This is because this algorithm can directly generate closed itemsets. Moreover, duplicated closed itemsets are detected and discarded from this algorithm.
Keywords :
data mining; query formulation; NCLOSED algorithm; best-first search strategy; itemset mining; Computer science; Data mining; Informatics; Information technology; Itemsets; Laboratories; Mathematics; Software algorithms; Software systems; Systems engineering and theory; Closed Itemset Mining; Data Mining;
Conference_Titel :
Convergence and Hybrid Information Technology, 2008. ICCIT '08. Third International Conference on
Conference_Location :
Busan
Print_ISBN :
978-0-7695-3407-7
DOI :
10.1109/ICCIT.2008.66
