Improving dual importance analysis based on a Shapley value associated with a fuzzy measure when interactions of criteria are significant

Kano's model is very useful to classify customer needs into different categories by completely using self-stated evaluations. However, the derived evaluation approach uses a less direct way of uncovering the evaluations that are most reliable to reflect the respondents' view from the survey. In addition, interaction effects among items, particularly non-linear interactions, are often incurred in practice. This study proposes a framework of using the dual importance graph with self-stated performance and derived importance computed by a Shapley value associated with a fuzzy measure method to classify the service items into different types of Kano's category by considering both linear and nonlinear effects among items. A case of evaluating the service quality of a particular hospital is illustrated to show how this proposed framework works. The result shows that using the Shapley value-based dual importance graph is more practical to deal with interactions of items.