Geetha, M and D'souza, R J (2008) Discovery of Frequent Closed Item sets using Reduced Pattern Count Tree. In: International Association of Engineers, 2-4 July, 2008, Imperial College London, London, U.K.

PDF Frequentclosed-1.pdf - Published Version Restricted to Registered users only Download (451kB) | Request a copy

Abstract

In this paper, a new algorithm for mining frequent closed itemsets from large volumes of data is implemented. A frequent itemset is maximal if none of its proper supersets is frequent. The total number of maximal frequent itemsets M is much smaller than that of frequent itemsets F, and we can derive each frequent itemset from M. However, M does not contain information of the support of each frequent itemset unless it is a maximal frequent itemset. Thus, mining only maximal frequent itemsets causes loss of information. However, when a transaction database is very dense and the minimum support is very low, i.e., when the database contains a significant number of large frequent itemsets, mining all frequent itemsets might not be a good idea. The concept of closed frequent itemsets solves this problem. This approach, uses a tree based data structure called Reduced Pattern Count Tree, and discovers all closed frequent itemsets in one scan of the database. On the other hand, the current algorithms need at least two scans of the database except Pattern Count Tree based algorithm, which requires a single scan of the database, but uses Lexicographical Ordered FP-tree to discover all frequent patterns.

Actions (login required)

View Item