2 WAYS FOR GETTING CONNECTIONS BETWEEN OBJECTIVE AND SUBJECTIVE DATA SETS IN MAN-MACHINE SYSTEMS - MULTIPLE CORRESPONDENCE FACTOR-ANALYSIS AND PROBABILISTIC FUZZY SET-THEORY

This paper suggests two approaches to establishing a connection between two sets of data, the first set considered as subjective variables, obtained from experts through continuous scale questionnaires, and the second set coming from indices measured on these individuals when performing a task. The first approach consists of using multiple correspondence factor analysis, in which all the variables are ''cut'' in fuzzy modalities in order to make data homogeneous and to bring to the fore nonlinear relationships. The second uses the probabilistic set theory in which data are synthesized by means of cumulative distribution functions and the relation between such functions is modeled using conditional possibility measures and fuzzy inference. Both methods give robust results and their differences are shown through a fictitious multidimensional example.