Research On Integration Of Structure Equation Model And Influence Diagrams In Customer Satisfaction Improving Decision

The integration of structure equation model (SEM) and influence diagrams (ID) is the main method and research object in this paper. After the two methods'characteristics are analyzed, respectively, the integration base and probability are studied, and customer satisfaction improving decision process is designed. The research content is outlined as follows.1.The research background, methods, ideas are described systematically. The research content and the integration of SEM and ID as the kernel method of customer satisfaction improving decision are confirmed.2.The customer satisfaction assessment of country and company is presented before fundamental and application of SEM are introduced. The technical and functional limitations of SEM are analyzed, which indicates that SEM is not capable of customer satisfaction improving decision.3.The customer loyalty is an important factor which connects customer satisfaction index and customer equity after theoretic analysis. Therefore the ID of customer satisfaction improving decision is designed which has value point used customer equity and decision point affecting basic distributions.4.The common base of SEM and ID in causality definition is studied. The detailed process, in which SEM and ID is integrated, is designed to make customer satisfaction improving decision. Methods are proposed to solve main technical problems, such as latent variable erasing, ID simplification ensured by coefficient or probability.5.The ID calculating process and computer program frame are designed, and relative formulas are described with emphasis on the kernel estimation which calculates distribution in the process. In view of the survey data, hypotheses about data explain capacity is established, and proper kernel function is selected.6.The customer satisfaction forecast process is proposed based on stable causality hypothesis. The simulating data are made by Monte Carlo with help of decision ID. Then the customer satisfaction index is forecasted by SEM using the data. The Monte Carlo simulation method and process are described in details.7.The ID calculation software is designed which is the application of theoretic research in SEM and ID integrating. The customer satisfaction improving decision of an airline company is studied. And the case results prove the feasibility and calculation validity of the integration of SEM and ID.