Research Of Customer Satisfaction Measurement Based On PLS Path Modeling

In this dissertation, based on the theory of Customer Satisfaction, Structural Equation Model, the method of Customer Satisfaction measurement, Customer Satisfaction regular model and expanded model are constructed. It shows the method to solve two models by PLS Algorithm and put forward several improving measures to Internet Survey. The processing of missing data has been taken into account adequately about parameters estimation of Structural Equation Model. A new method of missing data is put forward. The system of Customer Satisfaction measurement is completed.Main works of this dissertation is as follows:1. The research background, methods, goal, and main content are described systematically. The possibility and necessity of Internet Survey are confirmed about the research of Customer Satisfaction Index Measuring.2. The conception of Customer Satisfaction is summarized. The dissertation introduces the based theory of Structural Equation Model, presents the step of constructing Structural Equation Model, and describes the method of several Customer Satisfaction measurements.3. Four classical enterprise customer satisfaction index models and its structure characteristic are introduced. Wold's and Lohm?ller's PLS Algorithm are analyzed deeply. Customer Satisfaction measurement regular model and expanded model are put forward by comparing with Customer Satisfaction measurement model at home and aboard.4. The dissertation constructs Customer Satisfaction measurement regular model and expanded model. The setup, estimation and assessment of Customer Satisfaction regular model are analyzed mainly. It shows the method to evaluate the parameters of two models by PLS Algorithm. The advantage and disadvantage of Customer Satisfaction Internet Survey are analyzed deeply. The way of Internet Survey is improved. The missing data processing method of Weight Setup Newly (WSN) is presented. PLS Algorithm is finished based on Matlab.5. The design and implement of Customer Satisfaction measurement system are based on the theory and algorithm mentioned above. Customer Satisfaction measurement regular model and expanded model are analyzed and applied to evaluate food company and auto company respectively. The reliability, validity and goodness-of-fit index (GFI) of models are validated. The case results prove the feasibility and validity of the system. Compared the system with SmartPLS about astringency, weight coefficient and so forth, its iterative number is less and others are consistent. A new method and train is presented about theory research and application of Customer Satisfaction measurement in China.