Abstract

Generalization of itemsets of an organization with support and confidence metrics have helped the association rule mining by describing a preorder of transactions and their origin from the ancestral transactions. Intertransactions of different backgrounds stipulate far beyond metrics to be considered for mining process. Bounding the factors which determine the frequency of transactions, in turn simplifying the generations of candidate sets is the approach CBIT [Categorized and Bounded Inter-Transaction mining algorithm] to be discussed in this paper. Differentiation of intra and intertransactions lies in the limits applied to the customer, date and time or even a maintenance of records for a specific period of time. Intertransactions are limited by a very few of its ancestral constraints as they are intended to mine association rules from transactions and incitation of successive transactions of other domains. Hence the final methodology is adept in deriving the associative level of transactions in a bounded yet large domain of itemsets. Comparisons with previous renowned strategies would best describe the efficiency of this proposed technique.

Description