|
Along with the entry to the 21th century, comprehensive and profound transformation are happening in the world, international economic structure is being adjusted post, the trend to economic globalization is being accelerated day by day. Therefore, traditional measurement of microcosmic and microscopic quantity, such as output, throughput, GDP and so on, has been beyond not only the reflection to the competitive power of enterprise or industry, but also comprehensive national power of a country in world economic globalization competition;however, customer satisfaction with economic quality information can complement them effectively. Thus, scientific measurement of CSI is significant and pressing to improve the competition power of enterprise and measure national microscopic economic index. Meanwhile, the rapid development of computer technology, such as web technology、 database technology、 mathematics computation and program technology, and so on, has provided the support flat for the implement and expansion of CSI measure.At present, Professor fornell, the founder of American CSI proposed a measure method that uses PLS and SEM to measure CSI, which is very popular in the world. However, there exists some deficiency in the method, such as PLS can’t be ensured convergent, convergence speed is too slow, and so on, which need to research and resoluve. In the process of CSI measure, the lack of special measure software is another important problem, which impedes the comprehensive expansion and smooth process of CSI measure work seriously. Based on above background, under the support of National Science Fund item "Customer Satisfaction Index Model Computation", this paper extends the correlated research, and the content is arranged as following:First chapter elaborates the research background, the significance, the domestic and foreign research present situation of the topic. Second chapter introduces the related theory of CSI measure, including CSI model, fornell measure method, the related mathematic theory of measure, such as path analysis, SEM, PLS and so on. Third chapter further analyzes the question which exists in the SEM and the PLS method, and proposes our improvement program: Found the PLS iteration initialvalue under the unit vector restraint, enhance the PLS convergence rate, explain the PLS iteration convergence;Expand this kind of model and the algorithm to the multi-layer SEM;Found the steady and comprehensive CSI formula. Fourth chapter in detail introduces CSI measure software which are developed by us based on our improvement algorithm, including software design thought, system design, detailed design and realization and so on. Fifth chapter unifies the web technology % the data warehouse technology and so on, take the customer relations management (CRM) system as the reference, proposes the design of customer satisfaction analysis system. Sixth chapter has carried on the summary and the research forecast to the full text. The innovations of this dissertation can be showed in five aspects:(1) Find the PLS iteration initial value under the unit vector restraint, enhances the PLS convergence rate, explains the PLS iteration convergence.(2) Give multi-layer construction equation model and its algorithm.(3) Find a steady and comprehensive CSI final formula.(4) Realize the theory of evaluating CSI by programming, the measure software of CSI permit the diversification of investigation paper, and it has the contrast analysis function of correlation algorithm(5) Propose one design proposal of analysis system of customer satisfaction based on evaluating module.
|