Minimal Structure of Self-Organizing HCMAC Neural Network Classifier
Neural Processing Letters
Hi-index | 0.00 |
In this paper we describe a novel application of the CMAC (Cerebellar Model Articulation Controller) neural network. We have identified CMAC as an adaptive fuzzy system that can extract fuzzy rules from training data and use these rules for classifying a unknown input of a class. The extracted fuzzy rules are stored in CMAC's memory table and they do not need to be translated explicitly to be used for classification. We have also discovered that the CMAC's network operations are equivalent to fuzzy operations. This means that the CMAC, which has so far only been recognized as a controller and a function modeller, is in fact also a neuro-fuzzy recognition system.