IMPROVED ALGORITHM TO MINE ASSOCIATION RULES

Abstract

This work is an attempt to develop an efficient algorithm to mine association rules. The problem of association rule generation has recently gained considerable prominence in the data mining community because of its use as a tool for knowledge discovery. Consequently, there has been a spurt of research activity in the recent years surrounding this problem. Data mining is motivated by the decision support problem faced by most large retail organizations. Progress in bar-code technology has made it possible for retail organizations to collect and store massive amount of data, • referred to as the basket. data. A record in such data typically consists of the transaction data and the items bought in the transaction. Successful organizations view such databases as important component of the marketing strategy. They are interested in instituting information-driven marketing process, managed by database technology, which enables marketers to develop and implement customized marketing programs and strategies. An association rule identifies a combination of attribute or items that occur together with greater frequency than might be expected if the values or items were independent of one-another. Association rules find the relationship between the different attributes in a transaction database. Such rules track the patterns in transactions such as finding how the presence of one attribute in the transaction affects the presence of another and so forth. An association rule is the expression of the form A=>B where A and B are Boolean attributes and the symbol => is I called quantifier. The idea of an association rule is to develop a systematic method by which a user can figure out how to infer the presence of some sets of attributes, given the presence of other attributes in a transaction. Such information is useful in making decision such as customer targeting, shelving, and sales promotion. • Here the main focus is on reducing number of candidate item sets generated and number of database scans in the process of association rules mining.