A fuzzy filter for removing interferences among membership grade functions An application to pre-treatment of data in olive oil authentication

A fuzzy filter has been defined to partially eliminate interference between membership grade functions for the cases of absence and presence of uncertainty in input data. A functional study allowed us to analyse the effects on the qualified response of parameters involved in its definition, and hence choose the optimum values for them. Two applications concerning the authentication and sensory evaluation of olive oil are presented. Furthermore, the study reveals that its implementation in more general and sophisticated techniques is a straightforward matter, it being very appropriate for the pre-processing of fuzzy data.