Research On Customer Loyalty Based On Bayesian Network Approach

With the global economy integration process speeding up, the market competition is increasing. The attention on customer relationship management theory and the corresponding theoretical research is unprecedented. And customer loyalty directly decides the level of the enterprise to acquire the competitive edge in the market, so the theory of customer loyalty as a part of customer relationship management has become a hot spot.Compared with the traditional data mining methods, Bayesian network approach has significant advantages for graphical representation, clear causal relationship, supporting the reasoning on incomplete data and uncertainty reasoning, etc. This article applies Bayesian network approach in customer loyalty research.In this paper, we first analyze the status quo and problems about the research on theory of customer loyalty, and reviewed Bayesian network approach. Then we proposed to build a Bayesian network model with the combination of Bayesian Dirichlet scoring and Bound and Collapse method for customer loyalty and its related factors. The model can determine and analysis the relationship between customer loyalty and its related factors, which provided a scientific basis for taking effective measures to retain and acquire loyal customers. This paper gave the key steps to set up a Bayesian network model. And we gave a detailed description for the involved methods and problem. Finally, we did a case analysis with the data of customer service from some province branch of China Unicom, and the results proved the effectiveness of Bayesian network approach, which was put forward in this article.