Customer Behavior Modeling under Several Decision-Making Processes by using EM Algorithm

This paper gives a model of customer choice behavior modeling based on a combination of decision-making processes by applying latent class model based on EM algorithm. This model can apply for the choice problems of multi services and multi brands under various decision-making processes. In addition to the model based on EM algorithm, we tried some conventional models and compared them. The model based on EM algorithm enables us to know what kinds of customers are classified into a certain class. Moreover, we could construct more accuracy model than conventional model and found the existence of two decision-making processes.

[1]  Scott J. Savage,et al.  United States Demand for Internet Access , 2004 .

[2]  C. Manski.  The structure of random utility models , 1977 .

[3]  H. Kawano,et al.  Service demand analysis using multiattribute learning mechanisms , 2003, IEMC '03 Proceedings. Managing Technologically Driven Organizations: The Human Side of Innovation and Change (IEEE Cat. No.03CH37502).

[4]  A. Inoue,et al.  Service-demand-forecasting method using multiple data sources , 2006, Networks 2006. 12th International Telecommunications Network Strategy and Planning Symposium.

[5]  M. Ben-Akiva,et al.  Multi-featured products and services: analysing pricing and bundling strategies , 1998 .

[6]  Kenneth E. Train,et al.  Discrete Choice Methods with Simulation , 2016 .

[7]  R. Fildes,et al.  Telecommunications demand forecasting—a review , 2002 .

[8]  Takeshi Kurosawa,et al.  Telephone Service Choice-Behavior Modeling Using Menu Choice Data under Competitive Conditions , 2006 .