Design Procedure and Improvement of a Mathematical Modeling to Estimate Customer Satisfaction

Abstract

A design procedure of the customer satisfaction model, which is required for an artificial adaptive service system action, is introduced. And the methods to improve the model are presented using a statistical analysis and selection of several clustering methods that form the mathematical service model from the surveyed quantitative data. Using a case study about the hotel guest service, accuracies of those methods were evaluated by cross validations. As a result, the Naive Bayes clustering method and the REPTree algorithm showed good estimation of the customer satisfaction as much as about 40 %.