An efficient structure for fast mining high utility itemsets

Abstract

High utility itemset mining has emerged to be an important research issue in data mining since it has a wide range of real life applications. Although a number of algorithms have been proposed in recent years, there seems to be still a lack of efficient algorithms since these algorithms suffer from either the problem of low efficiency of calculating candidates' utilities or the problem of generating huge number of candidates. In this paper, we propose a novel data structure called PUNlist (PU-tree-Node list), which maintains both the utility information about an itemset and utility upper bound for facilitating the processing of mining high utility itemsets. Based on PUN-lists, we present a method, called MIP (Mining high utility Itemset using PUN-Lists), for fast mining high utility itemsets. The efficiency of MIP is achieved with three techniques. First, itemsets are represented by a highly condensed data structure, PUN-list, which avoids costly, repeatedly utility computation. Second, the utility of an itemset can be efficiently calculated by scanning the PUN-list of the itemset and the PUN-lists of long itemsets can be fast constructed by the PUN-lists of short itemsets. Third, by employing the utility upper bound lying in the PUN-lists as the pruning strategy, MIP directly discovers high utility itemsets from the search space, called set-enumeration tree, without generating numerous candidates. Extensive experiments on various synthetic and real datasets show that PUN-list is very effective since MIP is at least an order of magnitude faster than recently reported state-of-the-art algorithms on average.

Parallel Algorithm to Efficiently Mine High Utility Itemset

“…All itemsets that has TWU below the minimum utility threshold can be pruned, thus a complete HUIs can be discovered. This concept is used by many tree based algorithms [6,9,10]. CTU-PRO [6] has efficient tree structure called CUP-Tree.…”

“…That is the reason why HUIM is more complex thus designing efficient algorithm is very important. HUIM has been studied by many researchers, tree structure [6,9,10,17] and utility list based structure [7,[11][12][13][14] are the most interesting extensions because it has efficient data structure and also has efficient mining algorithm. Hybrid implementation [15,16] is also done to combine advantages from different data structures to improve the performance.…”

FCHUIM: Efficient Frequent and Closed High-Utility Itemsets Mining

“…In the search process of HUI-Miner, some information of items is stored in the utility list, and the form of the information is denoted as (tid, iutil, rutil). All the algorithms [18,19,21,23,25,[38][39][40] utilize a utility-list structure to mine itemsets in one phase. This is because, in the generation procedure, the utility lists of itemsets are joined to create new utility lists of their supersets without repeatedly scanning the database.…”

“…In addition, the precheck method is used to check whether the T idset of an itemset is contained by the T idset of its extensions and to determine whether the itemset is nonclosed (line 13). Then, the T SLs for extensions of a are recursively built (lines [17][18][19][20][21][22][23][24][25][26][27]. Then, FCHUIs are produced in the Generate-FCHUI procedure (line 30).…”

“…Therefore, HUI is determined as an itemset whose utility is not less than the minutil set by the user. HUIM [16][17][18][19][20][21][22][23][24][25] is an interesting research area, where researchers can focus on the values of items or the profits of commodities. HUIM is a difficult task because the utility measure used in HUIM is different from the support measure used in FIM; it is not antimonotonic.…”