- 學位論文
Applying Weighted Association Rules with the Consideration of Product Item Relevancy
Applying Weighted Association Rules with the Consideration of Product Item Relevancy
若您是本文的作者，可授權文章由華藝線上圖書館中協助推廣。
摘要
With the development of technology, companies apply data warehouse systems and use the tools of data mining to realize consumers’ needs. Association rule which is the one of data mining functions has been wildly used for several years. However, Apriori algorithm which is the most famous and original algorithm in data mining is not the flawless algorithm. Many researchers proposed improved algorithms to overcome the problems in Apriori, but none of these algorithms are always successful. In this study, we would like to introduce you a new concept of weighted association rule mining. The purpose is to discover cross section relationship among items and extract the unknown patterns such as diaper and beer. We proposed two algorithms called HWA (O) and HWA (P) and assign weights according to hierarchical methods. Hierarchical weights in HWA (O) are assigned by intuition; another is assigned by proportional thoughts. We compare performance of the number of large itemsets, number of rules, and content of rules in HWA (O), HWA (P), and Apriori algorithms. As the result, HWA (P) performs better than other algorithms. Then, we apply HWA (P) to a real world case study in retailing industry and we discover interesting patterns. At the end of the research, we provide suggestions for retailing managers based on the discovery in the study. To conclude, the algorithm we proposed can efficiently filter out the minor rules and extracts the implicit and unknown patterns. Not only retailing, the algorithm can be used in various industries. Marketing managers can also make decisions more precisely and satisfy customers’ needs at the same time.
並列摘要
With the development of technology, companies apply data warehouse systems and use the tools of data mining to realize consumers’ needs. Association rule which is the one of data mining functions has been wildly used for several years. However, Apriori algorithm which is the most famous and original algorithm in data mining is not the flawless algorithm. Many researchers proposed improved algorithms to overcome the problems in Apriori, but none of these algorithms are always successful. In this study, we would like to introduce you a new concept of weighted association rule mining. The purpose is to discover cross section relationship among items and extract the unknown patterns such as diaper and beer. We proposed two algorithms called HWA (O) and HWA (P) and assign weights according to hierarchical methods. Hierarchical weights in HWA (O) are assigned by intuition; another is assigned by proportional thoughts. We compare performance of the number of large itemsets, number of rules, and content of rules in HWA (O), HWA (P), and Apriori algorithms. As the result, HWA (P) performs better than other algorithms. Then, we apply HWA (P) to a real world case study in retailing industry and we discover interesting patterns. At the end of the research, we provide suggestions for retailing managers based on the discovery in the study. To conclude, the algorithm we proposed can efficiently filter out the minor rules and extracts the implicit and unknown patterns. Not only retailing, the algorithm can be used in various industries. Marketing managers can also make decisions more precisely and satisfy customers’ needs at the same time.
參考文獻
[1] Agrawal, R.,Imielinski, T.,Swami, A.,1993,"Mining Association Rules Between sets of items in large databases".In:Proceedings of the ACM SIGMOD Conference on Management of Data,pp207-216.
[6] Berry, M. J.A., Linoff, G.S., 1999, "Mastering Data Mining, The Art & Science of Customer Relationship Management", John Wiley & Sons, Inc..
[7] Berzal, F., Cubero, J.C., Martin, N., Serrano, J.M., “TBAR: An efficient method for association rule mining in relational databases”, Data & Knowledge Engineering, Vol. 37 , No.1, 2001, pp.47-64.
[8] Bose, I., Mahapatra, R.K., 2001, “Business data mining－a machine learning perspective”, Information and Management 39, pp.211-225
[13] Carven, M.W., and Shavlik, J.W., 1997, "Using Neural Networks for Data Mining", Future Generation Computer System, vol.13, pp.221-229.