Fuzzy information in knowledge representation and processing for frame-based structures

This paper presents selected aspects of information processing in frames. The slots of frames are considered as taking on linguistic values represented as fuzzy sets. Matching procedures are developed and algorithms resulting within this context are proposed. The question of modeling cases involving many exemplar frames associated with a single prototype frame is posed. Subsequently the resulting model is effectively applied to fill the values missing in an exemplar frame. The optimization task involving entropy criterion enables to minimize ambiguity associated with the reconstruction problem. The resulting interval-valued fuzzy constructs are interpreted in terms of relevancy of the reconstructed information.<
>