2002-A data mining based approach to a 3M problem in the context of electronic commerce

3M

A data mining based approach to a 3M problem in the context of electronic

commerce

NSC 91-2416-H-390-005

91

8

1

92

7

31

3M Many Many Many Many Many Many 3M ( ) ( )3M Abstract

In the relationship between a customer and a web store, often a 3M relationship is ignored. For a customer, the purchasing activity is like a 3M game where many customers want to buy many products provided by many web stores. For a web store, it also follows the 3M that many web stores want to sell many products to many customers. It is believed that the future is a function of the past. Consequently, knowledge discovered in databases will be very valuable to EC customers and merchants. In this research project, an application model is introduced that the technique of knowledge discovery is employed to help reveal two types of information that embrace customer purchasing characteristics and target customer. A practical data set in the context of insurance is used to demonstrate the proposed application model.

Keywords: KDD, DM, financial services industry

[1, 2, 3, 4, 6, 17, 18]

Haffman [1] Harvard Business Review IBM [8] [2, 5, 7, 10, 20, 22] Data Mining [9, 10, 11, 12, 13, 17, 19, 24, 26, 27]

(knowledge discovery in database)

[14, 15, 16, 21, 23]

…

Customer Relationship Management CRM

3M Many Many Many Many Many Many M Shaw [28]

Aha [14 ] (Bayesian Network)

AT&T (Response model)

Desarbo [15,16] CRISP Ahn [14]

AT&T Pitta[19] Kahan[21]

Jutkins[22]

1

/

)

log

(

....

)

log

(

)

,...

,

(

n 2 n 1 2 1 c c c c n c 2 c 1 c M M M M

n

n

n

n

n

n

n

I

−

+

+

−

=

(1)

n

_Ci The number of records that return to class Ci,

i=1,2,....,n.

M The total number of records.

)]

,....,

,

(

)

[(

)

(

2 1 1 m c v c v c v t i M vi

I

a

_i

a

_i

a

_i

n

A

E

=

=

(2)

t The number of different values that appear in attribute A.

nvi The total number of records that attribute A takes value Vi , i=1,2,...,t.

the total number of records that attribute A takes value Vi and returns to class Cj, i=1,2,...,t,

j=1,2,...,m.

M The total number of records.

)

(

)

,...

,

(

)

(

n c 2 c 1 c

n

n

E

A

n

I

A

Gain

−

=

(3) ID3 1.8087 (0.1476)
  
   
     283 95 1 7 24 100 ” ” 1. ” ” ( ) 2. ” 3.

[1] Hoffman, D. L. and Novak, T. P., "How to acquire customers on the web," Harvard Business Review, Vol. 78, Iss. 3, May.-Jun. 2000, pp. 179-183.

[2] Kannan, P. K., Chang, A. M., and Whinston, A. B., "Marketing information on the I-way," Communications of the ACM, Vol. 41, No. 3, Mar. 1998, pp. 35-43.

[3] Turban, E., Lee, J., King, D., and Chung, H. M., "Advertisement in electronic Commerce," in Electronic Commerce: A Managerial Perspective, Prentice-Hall, 2000.

[4] Allen, C., Kania D., and Yaeckel, B., Internet World Guide to One-To-One Web Marketing, John Wiley & Sons, 1998.

[5]

85 2

[6] Hesler, M., “Bridge the gap between online and off-line customer profiling and personalization,”

Target Marketing, Vol. 22, Issue 11, 1999, pp. 57. athttp://msdn.microsoft.com/workshop/manageme nt/planning/site052099.asp.

[7] Ochs, N. V., “Personalization and customization: where are they now?” Jun. 1999, at

[8]http://www-3.ibm.com/solutions/businessitelligence [9] Quintas, P. , Lefrere, P. and Jones, G. 1997 “Knowledge Management: A Strategic Agenda”, Long Range Planning , Vol. 30, No. 3, pp. 385-391.

[10] Manville and Foote, (1996) ‘Strategy as if knowledge mattered. Business Information: what is it really worth?”, Information Strategy, October, 1996.

[11] Bolisani, E. & Scarso, E. (1999), ‘Information technology management: a knowledge-based perspective’, Technovation, Vol. 19, No. 4, pp. 209-217.

[12] Quinlan, J.R. (1986), Induction of decision tree, Machine Learning, 1, pp.81-106.

[13]Breiman, L., Friedman, J.H., Olshen, R.A. & Stone, C.J. (1984), Classification and Regression Trees, Belmont CA: Wadsworth International Group.

[14]Ahn, Jae-Hyeon & Ezawa, Kazuo (1997), “Decision support for real-time telemarketing operations through Bayesian network learning”, Decision Support Systems, Vol. 21, pp.17-27. [15] Desarbo, W.S. & Ramaswamy, V. (1994),

“CRISP: customer response based iterative segmentation procedures for response modeling in direct marketing”, Journal of Direct Marketing, Vol. 8, No.3, pp.7-20

[16] Desarbo, W.S. & Ramaswamy, V. (1994), “Customer response based iterative segmentation procedures for response modeling in direct marketing”, Journal of Direct Marketing, Vol. 8, No.3, pp.7-20

[17] Han, J. (1998), “Towards on-line analytical mining in large databases”, SIGMOD, Vol. 27, No.1, pp.97-107.

[18] Galbreath, J. & Rogers, T. (1999), “Customer relationship leadership: a leadership and motivation model for the twenty century business”, The TQM Magazine, Vol. 11, No.3, pp.161-171.

[19] Pitta, D. (1998), “Marketing on-to-one and its dependence on knowledge discovery in databases”, Journal of consumer marketing, Vol. 15, No.5, pp. 468-480.

[20] Stone, M., Woodcock, N., & Wilson, M. (1996), “Managing the change from marketing planning to customer relation management”, Long Range Planning, Vol. 29, No. 5, pp. 675-683.

[21] Kahan, R. (1998), “Using database marketing techniques to enhance your one-to-one marketing initiatives”, Journal of Consumer Marketing, Vol.15, No.5, pp.491-493.

[22] Jutkins, R (1994), “Just image! Database marketing targets the right customers – and keeps them coming back”, Direct Marketing, Vol.12, No.

56, pp.38-40.

[23] Hui, S.C. & Jha, G. (2000), “Data mining for customer service support”, Information and Management, Vol.38, No. 1, pp.1-13.

[24]Brachman, R., Khabaza, T., Kloesgen, W., Piatestsky-Shapiro, G. & Simoudis, E. (1996), “Mining business databases”, Communication of the ACM, Vol. 39, No. 11, pp. 42-48.

[25] Walter, A. (1999), “Customer Relationship Promoters: Driving Forces for Successful Customer Relationships”, Industrial Marketing Management, Vol. 28, No. 5, pp. 537-551. [26] Chen, M.S., Han, J. & Yu, P.S. (1996), “Data

mining: an overview from a database perspective”, IEEE Trans. on Knowledge and Data Engineering, Vol. 8, pp.866-883.

[27] Fayyad, U. & Stolorz, P. (1997), “Data mining and KDD: promise and challenge”, Future Generation Computer Systems, Vol. 13, No. 2-3, pp.99-115.

[28] Shaw, MJ. Subramaniam, C., Tan, G.W., and Welge, M.E., (2001), “Knowledge management and data mining for marketing”, Decision Support Systes, Vol. 127-137.

1 ( ) 0~5000 5001~20000 20001~50000 50001~150000 150000 2 ( ) 1~5 6~15 15 3 ( ) 1~10 11~20 20 4 ( ) 0~14 15~30 31~50 51 5 ( ) 100 101~300 300~1000 1000 (class) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Rule 1 10 95 2 10 50 3 10 39 4 10 32 5 10 27 6 10 23 7 10 22 8 10 22 9 10 22 10 10 21 11 10 19 12 1 16 13 10 14 14 10 13 15 1 12 16 10 11 Rule 17 10 11 18 1 10 19 10 10 20 10 10 21 9 9 21 10 8 22 1 7 23 10 7 24 10 7 Transaction data Model mechanism Loading Mining Result Knowledge