Kurtosis-Based CRTRL Algorithms for Fully Connected Recurrent Neural Networks

Basit Öğe Kaydı
dc.authorid0000-0003-1186-3058en_US
dc.contributor.authorMenguc, Engin Cemal
dc.contributor.authorAcır, Nurettin
dc.date.accessioned2021-03-20T20:12:56Z
dc.date.available2021-03-20T20:12:56Z
dc.date.issued2018
dc.departmentBTÜ, Mühendislik ve Doğa Bilimleri Fakültesi, Elektrik Elektronik Mühendisliği Bölümüen_US
dc.descriptionMenguc, Engin Cemal/0000-0002-0619-549Xen_US
dc.description.abstractIn this paper, kurtosis-based complex-valued real-time recurrent learning (KCRTRL) and kurtosis-based augmented CRTRL (KACRTRL) algorithms are proposed for training fully connected recurrent neural networks (FCRNNs) in the complex domain. These algorithms are designed by minimizing the cost functions based on the kurtosis of a complex-valued error signal. The KCRTRL algorithm exploits the circularity properties of the complex-valued signals, and this algorithm not only provides a faster convergence rate but also results in a lower steady-state error. However, the KCRTRL algorithm is suboptimal in the processing of noncircular (NC) complex-valued signals. On the other hand, the KACRTRL algorithm contains a complete second-order information due to the augmented statistics, thus considerably improves the performance of the FCRNN in the processing of NC complex-valued signals. Simulation results on the one-step-ahead prediction problems show that the proposed KCRTRL algorithm significantly enhances the performance for only circular complex-valued signals, whereas the proposed KACRTRL algorithm provides more superior performance than existing algorithms for NC complex-valued signals in terms of the convergence rate and the steady-state error.en_US
dc.identifier.doi10.1109/TNNLS.2018.2826442en_US
dc.identifier.endpage6131en_US
dc.identifier.issn2162-237X
dc.identifier.issn2162-2388
dc.identifier.issue12en_US
dc.identifier.pmid29994052en_US
dc.identifier.scopusqualityQ1en_US
dc.identifier.startpage6123en_US
dc.identifier.urihttp://doi.org/10.1109/TNNLS.2018.2826442
dc.identifier.urihttps://hdl.handle.net/20.500.12885/745
dc.identifier.volume29en_US
dc.identifier.wosWOS:000451230100028en_US
dc.identifier.wosqualityQ1en_US
dc.indekslendigikaynakWeb of Scienceen_US
dc.indekslendigikaynakScopusen_US
dc.indekslendigikaynakPubMeden_US
dc.institutionauthorAcır, Nurettin
dc.language.isoenen_US
dc.publisherIeee-Inst Electrical Electronics Engineers Incen_US
dc.relation.ispartofIeee Transactions On Neural Networks And Learning Systemsen_US
dc.relation.publicationcategoryMakale - Uluslararası Hakemli Dergi - Kurum Öğretim Elemanıen_US
dc.rightsinfo:eu-repo/semantics/closedAccessen_US
dc.subjectAugmented statisticsen_US
dc.subjectcircular and noncircular (NC) complex-valued signalsen_US
dc.subjectkurtosisen_US
dc.subjectnonlinear complexvalued adaptive filteren_US
dc.titleKurtosis-Based CRTRL Algorithms for Fully Connected Recurrent Neural Networksen_US
dc.typeArticleen_US

Dosyalar

Koleksiyon

WoS İndeksli Yayınlar Koleksiyonu
Elektrik Elektronik Mühendisliği Bölümü Yayın Koleksiyonu
PubMed İndeksli Yayınlar Koleksiyonu
Scopus İndeksli Yayınlar Koleksiyonu