An on-line ICA-mixture-model-based fuzzy neural network

This work proposes a new fuzzy neutral network (FNN) capable of parameter self-adapting and structure self-constructing to acquire a small number of fuzzy rules for interpreting the embedded knowledge of a system from the given training data set. The propsed FNN is inherently a modified Takagi-Sugeno-Kang (TSK)-type fuzzy rule-based model with neural network´s learning ability. There are no rules initiated at the beginning and they are created and adapted through the newly propsed on-line independent component analysis (ICA) mixture model and back-propagation algorithm learning processing that performs simultaneous structure and parameter identification. Several experiments covering the areas of system identification and classification are carried out. These results show that the proposed FNN can achieve significant improvements in the convergence speed and prediction accuracy with simpler network structure.