Researcher:
Yılmaz, Yasin

Profile Picture
ORCID

Job Title

Master Student

First Name

Yasin

Last Name

Yılmaz

Name

Name Variants

Yılmaz, Yasin

Email Address

Birth Date

Filters

2009 - 20103

Advanced Search

Filter by

Settings

Researcher Of Publications: search.filters.isAuthorOfPublication.0d947793-4f29-4b33-81a5-0a31c05807a2

Search Results

Now showing 1 - 3 of 3
  • Placeholder
    PublicationMetadata only
    Competitive randomized nonlinear prediction under additive noise
    (IEEE-Inst Electrical Electronics Engineers Inc, 2010) N/A; Department of Electrical and Electronics Engineering; Department of Electrical and Electronics Engineering; Yılmaz, Yasin; Kozat, Süleyman Serdar; Master Student; Faculty Member; Graduate School of Sciences and Engineering; College of Engineering; N/A; 177972
    We consider sequential nonlinear prediction of a bounded, real-valued and deterministic signal from its noise-corrupted past samples in a competitive algorithm framework. We introduce a randomized algorithm based on context-trees [1]. The introduced algorithm asymptotically achieves the performance of the best piecewise affine model that can both select the best partition of the past observations space (from a doubly exponential number of possible partitions) and the affine model parameters based on the desired clean signal in hindsight. Although the performance measure including the loss function is defined with respect to the noise-free clean signal, the clean signal, its past samples or prediction errors are not available for training or constructing predictions. We demonstrate the performance of the introduced algorithm when applied to certain chaotic signals.
  • Placeholder
    PublicationMetadata only
    An extended version of the NLMF algorithm based on proportionate Krylov subspace projections
    (Ieee Computer Soc, 2009) N/A; Department of Electrical and Electronics Engineering; Department of Electrical and Electronics Engineering; Yılmaz, Yasin; Kozat, Süleyman Serdar; Master Student; Faculty Member; Graduate School of Sciences and Engineering; College of Engineering; N/A; 177972
    The Krylov proportionate normalized least mean square (KPNLMS) algorithm extended the use of proportional update idea of the PNLMS (proportionate normalized LMS) algorithm to the non-sparse (dispersive) systems. This paper deals with the mean fourth minimization of the error and proposes Krylov proportionate normalized least mean fourth algorithm (KPNLMF). First, the PNLMF (proportionate NLMF) algorithm is derived, then Krylov subspace projection technique is applied to the PNLMF algorithm to obtain the KPNLMF algorithm. While fully exploiting the fast convergence property of the PNLMF algorithm, the system to be identified does not need to be sparse in the KPNLMF algorithm due to the Krylov subspace projection technique. In our simulations, the KPNLMF algorithm converges faster than the KPNLMS algorithm when both algorithms converge to the same system mismatch value. The KPNLMF algorithm achieves this without any increase in the computational complexity. Further numerical examples comparing the KPNLMF with the NLMF and the KPNLMS algorithms support the fast convergence of the KPNLMF algorithm.
  • Placeholder
    PublicationMetadata only
    Competitive nonlinear prediction under additive noise
    (IEEE, 2010) N/A; Department of Electrical and Electronics Engineering; N/A; Department of Electrical and Electronics Engineering; Kozat, Süleyman Serdar; Yılmaz, Yasin; Faculty Member; Master Student; College of Engineering; Graduate School of Sciences and Engineering; 177972; N/A
    We consider sequential nonlinear prediction of a bounded, real-valued and deterministic signal from its noise-corrupted past samples in a competitive algorithm framework. We introduce a randomized algorithm based on context-trees [1]. The introduced algorithm asymptotically achieves the performance of the best piecewise affine model that can both select the best partition of the past observations space (from a doubly exponential number of possible partitions) and the affine model parameters based on the desired clean signal in hindsight. Although the performance measure including the loss function is defined with respect to the noise-free clean signal, the clean signal, its past samples or prediction errors are not available for training or constructing predictions. We demonstrate the performance of the introduced algorithm when its applied to certain chaotic signals.