Profit-based Association Rule Mining from Commercial Transactions

Abstract

Since various kinds of information systems have been introduced for managing market transactions, system managers can expect to apply information processing techniques for better performances (e.g., reducing costs, increasing profits and providing users with better services). To do so, many of the commerce systems have tried to apply association sale mining to efficiently conduct such management tasks. However, we have realized a problem on applying the existing association rule algorithms to the market transactions. Most of the rule mining schemes (e.g., Apriori algorithm) consider only the occurrences of the items by ignoring the profit of items. Thereby, we address a novel algorithm, profit-based association rule algorithm, to consider the relationship between the occurrence and the profit of each item during conducting data mining process. Through several experiments, we have shown that these optimization techniques can yield significant performance benefit.