Recursive fuzzy c-means clustering for recursive fuzzy identification of time-varying processes

Dovžan, Dejan; Škrjanc, Igor

doi:10.1016/j.isatra.2011.01.004

Cited by 88 publications

(38 citation statements)

References 32 publications

(41 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Therefore, the appliance of an initialization method is of great importance, in order to reduce the computational cost and increase the algorithm's performance. As an initialization method will be used the fuzzy c-means clustering algorithm [2,5].…”

Section: Fuzzy C-means Clusteringmentioning

confidence: 99%

Evolutionary fuzzy models for nonlinear identification

Mendes

Pinto

Araújo

et al. 2012

Proceedings of 2012 IEEE 17th International Conference on Emerging Technologies &Amp; Factory Automation (ETFA 2012)

View full text Add to dashboard Cite

This paper proposes a new method for identification problems for industrial applications based on a TakagiSugeno (T-S) fuzzy model. The learning of the T-S model is performed from input/output data to approximate unknown nonlinear processes by a coevolationary genetic algorithm (GA). The proposed method is an automatic tool since it does not require any prior knowledge concerning the structure (e.g. the number of rules) and the database (e.g. antecedent fuzzy sets) of the T-S fuzzy model, and concerning the selection of the adequate input variables and their respective time delays. The proposed methodology is able to design all the parts of the T-S fuzzy prediction model and it is composed by five hierarchical levels.To validate and demonstrate the performance and effectiveness of the proposed algorithm, it is applied on Box-Jenkins benchmark problem.

show abstract

Section: Fuzzy C-means Clusteringmentioning

confidence: 99%

Evolutionary fuzzy models for nonlinear identification

Mendes

Pinto

Araújo

et al. 2012

Proceedings of 2012 IEEE 17th International Conference on Emerging Technologies &Amp; Factory Automation (ETFA 2012)

View full text Add to dashboard Cite

show abstract

“…In this method, if there is no admission to the new fuzzy rule by input data, then only the parameters of the nearest rule are updated by using an extended Kalman filter (EKF) scheme. Dovzan and Skrjan [10] proposed an on-line TSK-type fuzzy model, which can be used for modeling control system or robotics by combination of a recursive fuzzy c-means and least squares. This method needs more computational cost than the SAFIS because of the fuzzy covariance matrix.…”

Section: Introductionmentioning

confidence: 99%

A new data-driven neural fuzzy system with collaborative fuzzy clustering mechanism

Prasad

Lin

et al. 2015

Neurocomputing

View full text Add to dashboard Cite

a b s t r a c tIn this paper, a novel fuzzy rule transfer mechanism for self-constructing neural fuzzy inference networks is being proposed. The features of the proposed method, termed data-driven neural fuzzy system with collaborative fuzzy clustering mechanism (DDNFS-CFCM) are; (1) Fuzzy rules are generated facilely by fuzzy c-means (FCM) and then adapted by the preprocessed collaborative fuzzy clustering (PCFC) technique, and (2) Structure and parameter learning are performed simultaneously without selecting the initial parameters. The DDNFS-CFCM can be applied to deal with big data problems by the virtue of the PCFC technique, which is capable of dealing with immense datasets while preserving the privacy and security of datasets. Initially, the entire dataset is organized into two individual datasets for the PCFC procedure, where each of the dataset is clustered separately. The knowledge of prototype variables (cluster centers) and the matrix of just one halve of the dataset through collaborative technique are deployed. The DDNFS-CFCM is able to achieve consistency in the presence of collective knowledge of the PCFC and boost the system modeling process by parameter learning ability of the self-constructing neural fuzzy inference networks (SONFIN). The proposed method outperforms other existing methods for time series prediction problems.

show abstract

“…The methods are based on learning algorithms for neural networks (Werbos, 1974), evolving clustering (Kasabov and Song, 2002), subtractive clustering (Angelov and Filev, 2004), fuzzy c-means clustering (Dovžan andŠkrjanc, 2011b), Gustafson-Kessel clustering (Dovžan andŠkrjanc, 2011a) and others (Johanyák and Papp, 2012;Vaščák, 2012;Rȃdac et al, 2011).…”

Section: Introductionmentioning

confidence: 99%

“…The problem of identification of dynamic systems is that with the local approach the local models have more appropriate local behaviour while the fuzzy model is less accurate globally (Yen et al, 1998;Sorensen, 1993 The second problem in nonlinear system identification is to properly partition the space of antecedent variables. The methods are based on learning algorithms for neural networks (Werbos, 1974), evolving clustering (Kasabov and Song, 2002), subtractive clustering (Angelov and Filev, 2004), fuzzy c-means clustering (Dovžan andŠkrjanc, 2011b), Gustafson-Kessel clustering (Dovžan andŠkrjanc, 2011a) and others (Johanyák and Papp, 2012;Vaščák, 2012;Rȃdac et al, 2011).…”

Section: Introductionmentioning

confidence: 99%

Comparison of Approaches for Identification of All-data Cloud-based Evolving Systems

2015

Self Cite

View full text Add to dashboard Cite

In this paper we deal with identification of nonlinear systems which are modelled by fuzzy rule-based models that do not assume fixed partitioning of the space of antecedent variables. We first present an alternative way of describing local density in the cloud-based evolving systems. The Mahalanobis distance among the data samples is used which leads to the density that is more suitable when the data are scattered around the input-output surface. All the algorithms for the identification of the cloud parameters are given in a recursive form which is necessary for the implementation of an evolving system. It is also shown that a simple linearised model can be obtained without identification of the consequent parameters. All the proposed algorithms are illustrated on a simple simulation model of a static system.

show abstract

Recursive fuzzy c-means clustering for recursive fuzzy identification of time-varying processes

Cited by 88 publications

References 32 publications

Evolutionary fuzzy models for nonlinear identification

Evolutionary fuzzy models for nonlinear identification

A new data-driven neural fuzzy system with collaborative fuzzy clustering mechanism

Comparison of Approaches for Identification of All-data Cloud-based Evolving Systems

Contact Info

Product

Resources

About