Customer Satisfaction Data mining Based on Fuzzy Technology

摘要

A new method of customer satisfaction (CS) data mining is proposed in detail. Based on sample analyzing, adopting the technology of fuzzy c-means (FCM), SVD etc., the hierarchical model of CS analysis is designed using the theory of hierarchical rule structure. In such model, the CS analysis system is partitioned into seven subsystems; each of them holds independent functions. After building initial rule base using FCM, the rule base with much fewer rules is created using fuzzy rule amplification. And then the rule base is optimized by SVD, interpolation and genetic algorithm. The result of test shows the CS analysis fuzzy system constructed in this paper possess good performance.