The Study On CSI Model Constructing & GME Integration Evaluating Of Its Parameter

Being the result of market economy competition and the embodiment of people's economic living welfare and an important intangible asset of enterprise, Customer satisfaction has become the standard of judging total product (service) quality, and that integrated CSI has been the mode of economy operation quality in new economy. This dissertation focuses on the study of CSI model construction and it's parameter estimation.The content of this dissertation is divided into six chapters. The first chapter outlines development of customer satisfaction concept, and reviews state of the research on customer satisfaction model, and analyses the implications of measuring customer satisfaction. The second chapter analyses the properties of SCSB, ACSI, ECSI and CCSI model and puts forward a few of suggestions of constructing CSI model, and points out advantages of measuring cumulative satisfaction through the annotation of customer satisfaction from an economic psychology perspective, and analyses the signification of customer loyalty in detail, and constructs a new CSI model on the basis of all-sided analysis on relationships between cause and effect of customer satisfaction. The third chapter analyses LISREL and PLS modeling principle, and indicates the necessity of integrating LISREL and PLS for measuring CSI. The forth chapter puts forward two methods of measuring CSI ( GME I and GMEII) based on the generalized maximum entropy(GME), and stresses on the analysis of GME regression principle, and gives the steps of determining weight by principal component analysis (PCA) and entropy value method (EV) in view of support space. The fifth chapter makes a comprehensive analysis on the actual CSI project, first estimating model parameters by LISREL, PLS, PCA and EV, then making more reliable model estimation by GME I and GME II, and educing Some valuable conclusions for enterprise managing. The sixth chapter makes a summary of this dissertation and research perspective.The innovations of this dissertation can be showed in four aspects:(1) On the basis of understanding on customer satisfaction by many scholars, this dissertation redefines customer satisfaction and gives the economic psychology explanation of cumulative customer satisfaction. This dissertation also studies customer loyalty in detail, and indicates that customer loyalty should contain four components.( 2 ) Four principles are put forward for constructing CSI model by studying typical national customer satisfaction index. On the basis of all-sided analysis on relationships between customer expectation perceived quality customer satisfaction customer value enterprise image and customer loyalty, a new dynamic CSI model is constructed.( 3 ) Through a comparative analysis on LISREL and PLS modeling, this dissertation points out the Advantage and deficiency on measuring CSI by LISREL and PLS, and puts forward two methods of measuring CSI based GME principle(called GME I and GME II ) on the purpose of achieving more reliable on estimating CSI model.(4) For the optimization of GME II , genetic algorithm is used in measuring CSI in order to aim best result. Particularly Several different estimation methods can be compared with each other by GME II subject to Some hypotheses.