臺灣博碩士論文加值系統

直接行銷是資料探勘上的一個重要的應用領域，為了提升某些產品的銷售和增加利潤，公司或組織會利用客戶的歷史資料，建立新客戶的選擇模式。傳統直接行銷的目標選擇演算法，都只著重於選擇出高回應機率的客戶。但這些方法所面臨的共同問題是未必選出有價值的客戶。為了解決此一問題，本論文提出一個以考慮客戶價值為主的關聯規則分類演算法，特別針對每個客戶所能提供的利潤來做分類，期望在不降低客戶回應率的前提下，達到提高總利潤的目的。經過實驗證實，我們的方法可以多找出一些高價值的客戶，有效的提高總利潤。

Direct marketing is an important application of the data mining. To acquire new customers, a company will use historical data about customers to build a model for selecting new customers. Conventional direct marketing uses response rate as the sole criterion for customer selection; that is, the company will select a potential customer for a marketing campaign based on the probability for the potential customer to respond. While this approach guarantees to render high response rate, it needs not guarantee to get high gross profit for the company. This is because some potential customer with high response rate may contribute only a little profit to the company. In this thesis, we proposed a new profit-based customer selection approach. This approach not only considers the customer response rate, but also considers the value of the customer to the company. Experiments show that the proposed approach can select high value customer so as to increase the gross profit of the company.

目錄
中文摘要 I
ABSTRACT II
誌謝 III
目錄 IV
圖表索引 VI
第1章 、緒論 1
1.1 研究背景 1
1.2 資料探勘的技術 2
1.3 研究動機與目的 5
1.4 論文架構 7
第2章 、文獻探討 8
2.1 相關理論與文獻 8
2.1.1 直接行銷 8
2.1.2 電子廣告郵件行銷 14
2.1.3 電子郵件的特性 15
2.2 相關論文 19
2.2.1 MetaCost演算法 20
2.2.2 Direct Cost-Sensitive演算法 22
2.2.3 Focused Association Rules 演算法 25
第3章 、研究方法 29
3.1 方法流程 30
3.2 研究方法說明 32
3.2.1 步驟說明 32
3.3 詳細做法與實例 39
3.3.1 訓練階段 39
3.3.2 測試階段 47
第4章 、研究成果與實驗 51
4.1 實驗的建置 51
4.1.1 資料描述 51
4.1.2 選擇重要的客戶屬性 52
4.1.3 規則的建立 53
4.2 實驗結果 54
4.3 綜合分析與討論 63
第5章 、結論與未來方向 65
參考文獻 67
作 者 簡 介 70

參考文獻
[1] B. Lent, A. Swami, J. Widom,” Clustering association rulesData Engineering” 1997. Proceedings. 13th International Conference on , 7-11 April 1997 Page(s): 220 -231
[2] B. Liu,W. Hsu, Y. Ma,”Integrating Classification and Association Rule Mining”,in:Proceedings of the Fourth International Conference on Knowledge Discovery and Data Mining(KDD-98),New York, pp.80-86,1998
[3] B. Zadrozny ,C. Elkan, ”Learning and making decisions when costs when costs and probabilities are both unknown”In SIGKDD, pp.204-213. SIGKDD, August 2001.
[4] D.Haughton,S.Oulabi,”Direct marketing modeling with CART and CHAID.” Jouranl of Direct Marketing, 7:16-26, 1993.
[5]　G.V.Kass,” An Explontory Technique for Investigating Large Quantities of Categeorical Data”, Applied Statistics, Vol.29(1980) pp.119-127
[6] H.P. Brondmo,& G. Morre.”The Engaged Customer:The New Rules of Interner Direct Marketing”,New York:Harper Business. (2000)
[7] Kdd-cup-98 dataset
　 http://kdd.ics.uci.edu/databases/kddcup98/kddcup98.html
[8] K. Wang, Y. Jiang, L. V.S. Lakshmanan. "Mining unexpected rules by pushing user dynamics", SIGKDD 2003, Washington D.C., August, 2003
[9] K. Wang, S. Zhou, J. M. S. Yeung, Q. Yang,”Mining Customer Value: From Association Rules to Direct Marketing.” ICDE 2003: 738-740
[10] J. Han, M. Kamber, Data mining :concepts and techniques, Morgan Kaufmann Publishers, San Francisco,pp.311-313(2001)
[11] J.M. Sousa, S. Madeira,U. Kaymak“Modeling charity donations using target selection for revenue maximization” Fuzzy Systems, 2003. FUZZ ''03. The 12th IEEE International Conference on , Volume: 1 , May 25- 28, 2003 ,Page(s): 654 -659
[12] J.R Bult and T.J Wansbeek,”Optimal selection for direct mail. ”Marketing Science,14:378-394,1993.
[13] J.Zahavi and N.Levin.”Applying neural networks to target marketing.”Journal of Direct Marketing”,11:2-22,1997.
[14] P. Domingos. “MetaCost:A General Method for Making Classifiers Cost-Sensitive”In KDD99,pages 155-164, KDD,August 1999.
[15] 　R. Agrawal, T. Imielinski, A. Swami: "Mining Associations between Sets of Items in Massive Databases", Proc. of the ACM-SIGMOD 1993 Int''l Conference on Management of Data, Washington D.C., May 1993, 207-216.
[16]　R. Brady, What really works in e-mail marketing.Target Marketing. 24(2),46-58 (2001).
[17] S.Madeira and J.M.Sousa.”Comparison of target selection methods in direct marketing.”In proceedings of 2nd European Symposium on Intelligent Technologies,Hybrid System and their implementation on Smart Adaptive Systems (EUNITE2002),Albufeira,Portugal,　September 2002.
[18] U. Kaymak.” Fuzzy target selection using RFM variables” IFSA World Congress and 20th NAFIPS International Conference, 2001. Joint 9th , Volume: 2 , 25-28 July 2001
[19] Y. Ma,B. Liu,C.K. Wong,P.S. Yu,S.M. Lee “Targeting the Right Student Using Data Mining” Proceedings of the sixth ACM SIGKDD international conference on Knowledge discovery and data mining, August 2000, Pages: 457 — 464
[20] Y. Ma,B. Liu,C.K. Wong, P. S. Yu: “Scoring the Data Using Association Rules.” Appl. Intell. 18(2): 119-135 (2003)
[21] W. Pijls and R. Potharst, ”Classification based upon Frequent Patterns ” , in: D.Lukoes and G.J.Williams(eds),Proceeding of the Symposium on the Application of Artifical Intelligence in Industry,pp 87-94.Deakin University, Geelong,2000
[22] Hans Reter Brondmo著陳正平譯，Email行銷基本教練:量身訂做的時代來了，台北市，臉譜，民國九十二年。
[23] 台灣創意力行銷協會著，厲害的！永遠是直效行銷，台北巿，新衛文化出版,民國八十六年。
[24] 江尻弘原著陳文彬譯，直接行銷技術，台北巿，前程企管，民國七十五年。
[25] 呂永和、張淑貞,”Mining Multilevel Association Rules in Large Databases ”,7th Conference on Artificial Intelligence and Applications,TAAI2002
[26] 樂以媛譯，直效行銷術，台北市，聯經，民國八十一年。