Hierarchical fuzzy integral stated preference method for Taiwan's broadband service market

The stated preference method (or conjoint analysis) has been a popular method for measuring buyer tradeoffs among multi-attribute products or services. In addition, the multi-nomial logit (MNL) model is a popular model for the stated preference method, although it relies on the assumption that its attributes have no correlations. However, in social problems, attributes are often correlated with each other. Recently, the Choquet integral has been used to solve non-additive problems. This study combines the Choquet integral and the stated preference method to propose the hierarchical fuzzy integral stated preference (HFISP) method, and then analyze the narrowband service users' preferences for broadband service. The results demonstrate that the hierarchical fuzzy integral stated preference method performs better than the partitioned fuzzy integral multi-nomial logit (PFIMNL) model. The HFISP method is effective and can be applied to deal with real life problems since it solves correlated attribute problem in discrete choice behavior.