A recurrent self-evolving fuzzy neural network with local feedbacks and its application to dynamic system processing

Juang, Chia‐Feng; Lin, Yin-Ku; Tu, Chiu-Chuan

doi:10.1016/j.fss.2010.04.006

Cited by 95 publications

(64 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The parameter values given here are the most effective parameter set for the ABC algorithm. The results obtained are compared with different fuzzy-neuro-based models taken from [20,21]. …”

Section: Simulation Resultsmentioning

confidence: 99%

Training ANFIS by using the artificial bee colony algorithm

Karaboğa¹,

Kaya²

2017

Turk J Elec Eng & Comp Sci

View full text Add to dashboard Cite

show abstract

“…The parameter values given here are the most effective parameter set for the ABC algorithm. The results obtained are compared with different fuzzy-neuro-based models taken from [20,21]. …”

Section: Simulation Resultsmentioning

confidence: 99%

Training ANFIS by using the artificial bee colony algorithm

Karaboğa¹,

Kaya²

2017

Turk J Elec Eng & Comp Sci

View full text Add to dashboard Cite

show abstract

“…Our results demonstrate that MRIT2NFS can effectively capture information about the system using mutual feedbacks and outperforms the RSEIT2FNN, which only uses local feedbacks. The recurrent type-1 FNNs considered include the wavelet-based RFNN (WRFNN) [6], TSK-type recurrent fuzzy network with supervised learning (TRFN-S) [9], and RSEFNN-LF [11]. Table I indicates that the test error of the TRFN and the RIFNN are very close for the noise-free environment, but for noisy environment, the TRFN-S is found to perform better.…”

Section: A Example 1 (Dynamic System Identification)mentioning

confidence: 99%

“…Recurrent self-organizing neural fuzzy inference network [7] and Takagi-Sugeno-Kang (TSK)-type recurrent fuzzy network (TRFN) [9], [10] use a global feedback structure, where the firing strengths of all rules are summed and fed back as internal network inputs. The recurrent self-evolving fuzzy neural network with local feedbacks (RSEFNN-LF) is proposed in [11] for dynamic system identification, where the recurrent firing values are influenced by both prior and current values. All of the recurrent FNNs that we have discussed so far use type-1 fuzzy sets.…”

Section: Introductionmentioning

confidence: 99%

“…On the other hand, the consequent part in the RIFNN [38] is of Mamdani type, which is formulated as an interval-valued fuzzy set. In many papers, it has been seen that the TSK-type modeling can do an excellent job of modeling dynamic systems [9], [11], [30], [34], [36], [40]. This paper proposes a mutually recurrent interval type-2 neural fuzzy system (MRIT2NFS) for dynamic system identification.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Mutually Recurrent Interval Type-2 Neural Fuzzy System (MRIT2NFS) With Self-Evolving Structure and Parameters

Lin

Chang

Pal

et al. 2013

IEEE Trans. Fuzzy Syst.

View full text Add to dashboard Cite

In this paper, a mutually recurrent interval type-2 neural fuzzy system (MRIT2NFS) is proposed for the identification of nonlinear and time-varying systems. The MRIT2NFS uses type-2 fuzzy sets in order to enhance noise tolerance of the system. In the MRIT2NFS, the antecedent part of each recurrent fuzzy rule is defined using interval type-2 fuzzy sets, and the consequent part is of the Takagi-Sugeno-Kang type with interval weights. The antecedent part of MRIT2NFS forms a local internal feedback and interaction loop by feeding the rule firing strength of each rule to others including itself. The consequent is a linear combination of exogenous input variables. The learning of MRIT2NFS starts with an empty rule base and all rules are learned online via structure and parameter learning. The structure learning of MRIT2NFS uses online type-2 fuzzy clustering. For parameter learning, the consequent part parameters are tuned by rule-ordered Kalman filter algorithm to reinforce parameter learning ability. The type-2 fuzzy sets in the antecedent and weights representing the mutual feedback are learned by the gradient descent algorithm. After the training, a weight-elimination scheme eliminates feedback connections that do not have much effect on the network behavior. This method can efficiently remove redundant recurrence and interaction weights. Finally, the MRIT2NFS is used for system identification under both noise-free and noisy environments. For this, we consider both time series prediction and nonlinear plant modeling. Compared with type-1 recurrent fuzzy neural networks, simulation results show that our approach produces smaller root-meansquared errors using the same number of iterations.Index Terms-Dynamic system identification, on-line fuzzy clustering, recurrent neural fuzzy systems, type-2 fuzzy systems.

show abstract

“…The idea behind evolving fuzzy systems is to use the available online data for updating the current rule base. Evolving fuzzy systems have been implemented by combining fuzzy inference systems with neural networks [11], [12], [13], [14]. On the other hand, an unsupervised and non-iterative evolving fuzzy technique has also been introduced [4], [7], [15], [16], [17].…”

Section: A Evolving Fuzzy Systemsmentioning

confidence: 99%

Image classification using evolving fuzzy inference systems

Othman

Tizhoosh

2013

2013 Joint IFSA World Congress and NAFIPS Annual Meeting (IFSA/NAFIPS)

View full text Add to dashboard Cite

Evolving fuzzy systems change by online updating of their parameters and structure; the number of fuzzy rules changes as long as there is new data. In literature, an evolving fuzzy system is mainly considered to be an unsupervised approach that builds and updates its clusters online as long as new data is available. In our previous works, we introduced a new supervised evolving fuzzy approach for segmenting medical images. In this paper, we demonstrate that this supervised evolving fuzzy approach can classify images. As an example we attempt to classify medical images based on their modalities. A set of features extracted from the image is used to train the fuzzy system with the modality class of the image as the fuzzy output. The proposed algorithm is applied to both ultrasound scans and magnetic reasoning images (MRI). The proposed algorithm is compared with the support vector machines (SVMs) and the K-nearest neighbour algorithm (KNN). The results show that evolving fuzzy systems can compete with well-establish clustering algorithms (and even surpass them) by delivering high classification rates.

show abstract

A recurrent self-evolving fuzzy neural network with local feedbacks and its application to dynamic system processing

Cited by 95 publications

References 22 publications

Training ANFIS by using the artificial bee colony algorithm

Training ANFIS by using the artificial bee colony algorithm

A Mutually Recurrent Interval Type-2 Neural Fuzzy System (MRIT2NFS) With Self-Evolving Structure and Parameters

Image classification using evolving fuzzy inference systems

Contact Info

Product

Resources

About