A cognitive model for context dependent fuzzy knowledge

• Main Authors: Eksioglu, K.M., Lachiver, G.
• Format: Conference Proceeding
• Language: English
• Subjects: Context modeling; Fuzzy logic; Fuzzy systems; Information processing; Knowledge representation; Power system modeling; Psychology; System software; Temperature; Upper bound
• DOI: 10.1109/NAFIPS.1999.781732

Context dependency is an important problem of knowledge. Fuzzy information also suffers from this dependency when any fuzzy concept, once described, is isolated from its context. The paper describes a cognitive model in which fuzzy knowledge is processed with respect to its surrounding context. The model is derived from some psycho-cognitive approaches: the integrity of knowledge and its context as well as the episodic character of contextual effects on knowledge. The model consists of some functional modules performing related cognitive tasks. Fuzzy knowledge is kept within a layered long-term memory. Layers form an episodic base for contextual knowledge: each layer represents a different context and keeps related fuzzy knowledge. A context selector module evaluates existence of context information in order to choose appropriate context layers. An aggregation module which is an optimization module evaluates implicit contexts. The model is supported by a car brake example, which shows its simplicity and adaptability, and thus the power of this cognitive model. The model is useful especially for mobile fuzzy system applications where knowledge on environment conditions are constantly changing and adaptation of the fuzzy system's descriptive aspects are necessary.