在少樣商品或短交易長度情況下挖掘關聯規則

Mining Association Rules When Number of Items is Limited or Transaction Length is Short

10.6382/JIM.200301.0055

陳家仁(Jia-Ren Chen)；陳彥良(Yen-Liang Chen)；陳禹辰(Yu-Chen Chen)

資料挖掘 ； 關聯規則 ； 交易資料庫 ； Data mining ； Association rule ； Transaction database

資訊管理學報

9卷2期（2003 / 01 / 01）

55 - 72

繁體中文

從交易資料庫中挖掘出的關聯規則可以幫助組織實行目標行銷，如進行市場區隔、選擇目標顧客、改進賣場陳設或組合搭售商品。以往有關的研究大多假設在單一商店的商品項目即可能達到數萬種以上，同時顧客可能會同時採購非常多樣化的商品。但在實際的世界中，許多商店如專賣店、精品店、速食店、餐廳、保險公司、百貨公司中的專櫃等等，所販售的商品種類可能只有數十至數百種不到；此外，一般消費者在多數的商店中每次購買的商品的種類通常也不會太多。基於上述兩種情況，本文發展一個全新的挖掘關聯規則作法，針對挖掘關聯規則時最耗時的步驟加以改進，在掃瞄資料庫一次後，將資料庫的內容儲存於一個樹狀結構中，再利用此樹狀結構產生關聯規則。如此將可大幅減少I/○的時間，讓使用者能更快產生關聯規則，並且不需在掃瞄資料庫前即指定minimum support，可以動態給定minimum support而不用重新掃瞄資料庫。

The problem of mining association rules is to find the associations between items in a large database of sales transactions. Basically, the past researches studied the problem with the assumptions that a great number of different items are sold in a store and a customer may buy quite a few items in a single round of purchase. No doubt, such situations fit in with the retailing store or convenience store well. However, there are many situations in practice that only a limited number of items are sold or the average transaction length is short. The possible examples include shopping in luxury goods stores, electric appliance stores, musical instrument stores, cigar stores, wine stores, glasses stores, watch stores, make up stores, underwear stores and so on. In view of this difference, this paper develops a new algorithm for mining association rules in such a special situation: small transaction length and hundreds of different items. Our experiments show that the developed algorithm outperforms the currently best algorithm, PP tree algorithm, designed for mining association rules in general situations.

1. 沈清正 Shen, Ching-Cheng、陳仕昇 Chen, Shih-Sheng、高鴻斌 Gao, Hong-Bin、張元哲 Chang, Yuan-Zhe、陳家仁 Chen, Jia-Ren、黃琮盛 Huang, Cong-Sheng、陳彥良 Chen, Yen-Liang(2002)。資料間隱含關係的挖掘與展望 An Overview on Mining Implicit Data Relation。資訊管理學報 Journal of Information Management，9(專刊)
   連結：
2. 陳彥良 Chen, Yen-Liang、凌俊青 Ling, Chun-Ching、許秉瑜 Hsu, Ping-Yu(2001)。在包裹資料庫中挖掘數量關聯規則 Mining Quantitative Association Rules in Bag Databases。資訊管理學報 Journal of Information Management，7(2)
   連結：
3. Agrawal, R.,Srikant, R.(1994).Proceedings of the 20th International Conference on Very Large Databases.Santiago:Chile.
4. Chen, M. S.,Han, J.,Yu, P. S.(1996).Data Mining: An Overview from a Database Perspective.IEEE Transactions on Knowledge and Data Engineering,8(6)
5. Han, J.,Kamber, M.(2001).Data Mining: Concepts and Techniques.Academic Press.
6. Han, J.,Pei, J.,Yin, Y.(2000).Proceedings of the 2000 ACM-SIGMOD International Conference Management of Data (SIGMOD'00).Dallas, TX:
7. Park, Jong Soo,Chen, Ming-Syan,Yu, Philip S.(1997).Using a Hash-Based Method with Transaction Trimming for Mining Association Rules.IEEE Transactions on Knowledge and Data Engineering,9(5)
8. Pasquier, N.,Bastide, Y.,Taouil, R.,Lakhal, L.(1999).Efficient Mining of Association Rules using Closed Itemset Lattices.Information Systems,24(1)
9. Savasere, A.,Omiecinski, E.,Navathe, S.(1995).Proceedings of the 21th International Conference on Very Large Data Bases (VLDB).Zurich, Switzerland:
10. Shen, Li,Shen, Hong,Cheng, Ling(1999).New Algorithms for Efficient Mining of Association Rules.Information Sciences,118(1-4)
11. Toivonen, H.(1996).Proceedings of the 22th International Conference on Very Large Databases (VLDB'96).Mumbay, India:

1. Lin, Jan-Yan,Lin, Angel,Hu, Yi-Chung,Lin, Jan-Yan,Lin, Angel,Hu, Yi-Chung(2013).Analyzing Investment Regions in Mainland China for Taiwanese Firms by Association Rule Mining.Asia Pacific Management Review,18(2),143-160.
2. 胡宜中、林震岩、林雅惠(2011)。運用關聯規則和序列型樣探討投資地區之關聯性與遷移─以印刷電路板產業為例。明新學報，37(1)，217-230。
3. 簡禎富、楊景晴、彭金堂、張盛鴻(2005)。建構關聯規則資料挖礦架構及其在台電配電事故定位之研究。資訊管理學報，12(4)，121-141。