Discovering consumer's purchasing behavior based on efficient association rules
Chong Wang, Yanqing Wang · 2011
Mining generalized association rules between items in the presence of taxonomies has been recognized as an important model in data mining. The classic Apriori itemset generation works in the presence of taxonomy but fails in the case of nonuniform minimum supports. In this paper, we extended the scope of mining generalized association rules in the presence of taxonomies to allow any form of user-specified multiple minimum supports. This method considers taxonomy of itemset, and can discover some deviations or exceptions that are more interesting but much less supported than general trends. Finally, the algorithms is validated by the example of transaction database. The result indicates this algorithm is successful in discovering consumer's purchasing behavior by user specifing different minimum support for different items.