Descriptive and Comparative Analysis of Human Perceptions expressed through Fuzzy Rating Scale-based Questionnaires

Opinion surveys are widely admitted as a valuable source of information which becomes complementary to the information extracted from data by machine learning techniques. This paper focuses on a challenging and still open problem which is related to how to handle properly the inherent uncertainty of human perceptions. Namely, we propose new ways to interpret and analyze fuzzy data coming out from a special case of survey, the so-called fuzzy rating scale-based questionnaire. This kind of questionnaire is characterized by allowing expressing human perceptions in terms of fuzzy rating scales. The proposed methods are in charge of capturing and modeling the uncertainty of the answers by varying the heights of the related fuzzy sets. These methods have been validated in two case studies: (1) a descriptive survey related to the packaging design of gin bottles; and (2) a comparative survey related to 2015 IFSA-EUSFLAT conference.

keywords: data science, descriptive and comparative surveys, fuzzy rating scales, human perceptions