Mining frequent closed itemsets based on projection Array and closed itemsets lattice

Frequent closed itemsets are a lossless, condensed representation of all the frequent itemsets. Although many related algorithms have been suggested, the efficiency of these algorithms is not very high when the dataset is very dense. To address this problem, a novel algorithm called FCIL-Mine is proposed in this paper. Firstly, the concept of PArray is described, which is used for discovering those items that always appear together, and the corresponding algorithm is presented by using data horizontally and vertically. Then, a new data structure FCIL is proposed based on PArray, which is used for maintaining the essential information of frequent closed itemsets. Finally, all of the frequent closed itemsets can be found by using hash table check and subsume-check pruning strategy. The experiments show that FCIL-Mine is superior to DCI-Close in execution efficiency, especially when the datasets are very dense.