A Research Of Customer Classification Method Based On Granular Computing Theory

Nowadays, customers have become the most valuable resource for companies, therefore most companies take the customer relationship management (CRM) as the first place. The essence of customer relationship management is grasping the characteristics of customers in various aspects, thus companies provide the required core interests for customers. Through customer relationship management, companies can improve customer loyalty, and realize customer lifetime value of consumption. However, customers characteristics identification base on scientific classification. Due to the features of uncertainty, completeness and fuzzing about customer classification indicators, thus the traditional artificial intelligence method and hardly realize the precise customer classification reaults. In recent years, domestic scholars namely professor Zhang Bo and ZhangLing puts forward granular computing methods based on quotient space, the theory is good at solving uncertain problems. So, this paper introduces granular computing methods based on quotient space into customer classification area, aiming to provide reference for customer classification practice and theories research. This paper contains three contents:(1) Customer relationship management and customer classification methods (chapter 2). This chapter firstly expounds the connotation of customer relationship management and its importance. Subsequently, the chapter summarizes the existing customer classification methods from two aspects. The first one is based on customer characteristics, such as the indicators of the demographic variables, lifestyle and behavior, and put forward the customer loyalty classification method from the comprehensive features of three characteristics about customer above. The second aspect is based on the customers'classification methods, and summarizes the research status.(2) Granular computing theory and the principle for customer classification granularity theory (chapter 3). This part firstly expounds the definition and basic structure of granular computing, as well as three model of exiting granular computing theory. Finally, this part puts forward the principle about granular computing in customer classification. The researches above supplies the foundation for customer classification based on granular computing in chapter four.(3) Customer classification based on granularity theory (chapter 4). This part first proposed the definition and two kinds of indicators system of customer loyalty. Secondly, this part puts forward two different customer rough classification method based on customer's psychology and key behaviors. Thirdly, this part researches the principle, processes and procedures of customer loyalty classification based on granular computing theory. This part also verifies the reliability of the customer loyalty classification method with the enterprise customer information.Through the researches above, it's proved that granular computing based on quotient space in application of customer loyalty classification has good reliability and operability, and it is a useful classification method both for future theoretical research and companies'customer relationship management practice.