Processing of incomplete fuzzy data using artificial neural networks

A degenerated fuzzy-number processing system based on artificial neural networks (ANNs) is introduced. The digital representation of fuzzy numbers is assumed, where the universe of discourse is discretized into n equally divided intervals. The representation of the membership function values is transformed into binary quantized values which have their maximum at 2^m-1 where m is the number of data bits used in the system. It is proposed that fuzzy number processing be performed in two basic stages. The first is the retrieval of fuzzy data consisting of degenerated fuzzy numbers, and the second is the performance of the desired fuzzy operations on the retrieved data. A method of incomplete fuzzy-number retrieval is proposed. It is based on an ANN structure which is trained to estimate the missing membership function values