This is a collection of four different S-function implementations of the recurrent fuzzy neural network (RFNN) described in detail in . It is a four-layer, neuro-fuzzy network trained exclusively by error backpropagation at layers 2 and 4. The network employs 4 sets of adjustable parameters. In Layer 2: mean[i,j], sigma[i,j] and Theta[i,j] and in Layer 4: Weights w4[m,j]. The network uses considerably less adjustable parameters than ANFIS/CANFIS and therefore, its training is generally faster. This makes it ideal for on-line learning/operation. Also, its approximating/mapping power is increased due to the employment of dynamic elements within Layer 2. Scatter-type and Grid-type methods are selected for input space partitioning.
 C.-H. Lee, C.-C. Teng, Identification and Control of Dynamic Systems Using Recurrent Fuzzy Neural Networks, IEEE Transactions on Fuzzy Systems, vol.8, No.4, pp.349-366, Aug. 2000.
Ilias Konsoulas (2023). Recurrent Fuzzy Neural Network (RFNN) Library for Simulink (https://www.mathworks.com/matlabcentral/fileexchange/43021-recurrent-fuzzy-neural-network-rfnn-library-for-simulink), MATLAB Central File Exchange. Retrieved .
MATLAB Release Compatibility
Platform CompatibilityWindows macOS Linux
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!Start Hunting!
I have killed some redundant variables and commands. The new s-functions are more concise and therefore, easily readable. Naturally, faster execution should come as a result.
Minor corrections in the description of this submission.
Added some details in the Description entru of this form.