Neurofuzzy and EUBAFES as tools for knowledge discovery in visual field data

A knowledge based glaucoma monitor based on papilla findings, intraocular pressure and visual field data has been developed successfully. To explain and interpret the classifier decision concerning the visual field data in the single decision and over time a neurofuzzy approach has been applied. A distance based feature selection approach is used to find a feature subset which leads to well separated classes in the measurement space. It is also used as first step to get clusters of interest for neurofuzzy interpretation.