MINING ASSOCIATION RULES TO EVALUATE CONSUMER PERCEPTION: A NEW FP-TREE APPROACH

Abstract

Association rule mining finds interesting relationships among large set of data items. While finding the important (or, frequent) relations from the set of consumer survey data, a modified algorithm based on frequent pattern growth is developed in this work. The sensitivity of support and confidence used for rule mining on the data is tested. The interaction between the order of the attributes and the confidence used is observed in terms of the number of rules mined. Th e impact of the product features on the level of consumer perception is thoroughly studied.

Keywords

• Association rule mi ning
• Itemset
• Frequent pattern tree
• Support
• Confid ence
• Order
• Cons umer satisfaction
• Data mining
• Confidence
• Consumer satisfaction