Customer Satisfaction Measurement Model System

With the dramatic development of e-commerce, the competition of the global is becoming more and more extensive, and the competition rules between enterprises have changed deeply. Many enterprises adopt CRM and customer satisfaction to improve core competitive ability. The customer satisfaction theory is becoming the hot spot of enterprise management.Academic world have done a lot of research on CSI for a long time, but most of them pay more attention on the architecture and the method of CSI evaluation. As so far, the model of the CSI evaluation mainly is the ACSI, which adopt the PLS recursive structure equation. Some researchers have brought forward the fuzzy comprehensive evaluation method. When there is too many evaluation targets and ingredients, the above methods will lead to information loss and sharply accuracy decrease.This paper aims to the problems exist in the CSI model, Introduces a new theory, which utilize the strong no-liner mapping ability of artificial neural network. This new method can find the rules underlying the multi-factor data, and construct a CSI model based on artificial self-organize neural network, which mapping input data to a output pattern, and those output node can self-organize. With a number of practices, the new model can generate correct mapping. The neural network remembers a customer feeling pattern based on the input data, and can produces complex association and remembrance on customer satisfaction and carries the point to evaluating and analyzing the CSI.In this paper, we use J2EE technology to implement a customer satisfaction evaluation system, which proved the effective of the new model. This system has four function modules: providing data, statistics and analyzing, constructing model (training data and testing data), forecasting CS and analyzing factor. The system has friendly interface and excellent performance ,moreover, it has good universality and practicality of dealing with questionnaire data and constructing CSI model based on artificial self-organized neural network .