JISE


  [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15]


Journal of Information Science and Engineering, Vol. 33 No. 1, pp. 199-216


An Efficient Method for Mining Frequent Weighted Closed Itemsets from Weighted Item Transaction Databases


BAY VO1,2
1
Division of Data Science
Ton Duc Thang University
Ho Chi Minh City, 70000 Vietnam

2Faculty of Information Technology
Ton Duc Thang University
Ho Chi Minh City, 70000 Vietnam
E-mail: vodinhbay@tdt.edu.vn; bayvodinh@gmail.com


    In this paper, a method for mining frequent weighed closed itemsets (FWCIs) from weighted item transaction databases is proposed. The motivation for FWCIs is that frequent weighted itemset mining, as frequent itemset (FI) mining, typically results in a substantial number of rules, which hinders simple interpretation or comprehension. Furthermore, in many applications, the generated rule set often contains many redundant rules. The inspiration for FWCIs is that one potential solution to the rule interpretation problem is to adopt frequent closed itemset. This study first proposes two theorems and a corollary. One theorem is used for checking non-closed itemsets while joining two itemsets to create a new itemset and the other theorem is used for checking whether a new itemset is non-closed itemset or not. The corollary is used for checking non-closed itemsets when using Diffsets. Based on these theorems and corollary, an algorithm for mining FWCIs is proposed. A Diffset-based strategy for the efficient computation of the weighted supports of itemsets is described. A complete evaluation of the proposed algorithm is presented. 


Keywords: frequent weighted closed itemset, frequent weighted support, weighted itemset-tidset (WIT) trees, data mining, non-redundant rule

  Retrieve PDF document (JISE_201701_12.pdf)