An architecture of neural network for fuzzy teaching inputs
H.-M. Lee, Wun-Hao Wang · 2002
A neural network for classification problems with linguistic terms is proposed. A fuzzy input is represented as a LR-type fuzzy set. A generalized pocket algorithm, called a fuzzy pocket algorithm, that utilizes LR-type fuzzy sets operations and a defuzzification method is first applied to train a linear threshold unit (LTU). This LTU node will classify as many fuzzy input instances as possible. Afterwards, FV nodes that represent fuzzy vectors will then be generated and expanded by the FVGE learning algorithm to classify those input instances that cannot be classified by the LTU node. The network structure is automatically generated. Online learning is supplied, and the learning speed is fast. One sample problem, called a knowledge-based evaluator, is considered to illustrate the working of the proposed method. Also, the experimental results are very encouraging.