Detail publikace

Piecewise-polynomial curve fitting using group sparsity

Originální název

Piecewise-polynomial curve fitting using group sparsity

Anglický název

Piecewise-polynomial curve fitting using group sparsity

Jazyk

en

Originální abstrakt

We present a method for segmenting onedimensional signal, whose segments are modeled as polynomials, and which is corrupted by an additive noise. The method is based on sparse modeling and its formulation as a convex optimization problem is solved by proximal algorithms. We perform experiments on simulated data, discuss the results and suggest future directions that could lead to even better results.

Anglický abstrakt

We present a method for segmenting onedimensional signal, whose segments are modeled as polynomials, and which is corrupted by an additive noise. The method is based on sparse modeling and its formulation as a convex optimization problem is solved by proximal algorithms. We perform experiments on simulated data, discuss the results and suggest future directions that could lead to even better results.

Dokumenty

BibTex


@inproceedings{BUT127475,
  author="Michaela {Novosadová} and Pavel {Rajmic}",
  title="Piecewise-polynomial curve fitting using group sparsity",
  annote="We present a method for segmenting onedimensional signal, whose segments are modeled as polynomials, and which is corrupted by an additive noise. The method is based on sparse modeling and its formulation as a convex optimization problem  is solved by proximal algorithms. We perform experiments on simulated data, discuss the results and suggest future directions that could lead to even better results.",
  address="EDAS Conference Services",
  booktitle="Proceedings of the 8th International Congress on Ultra Modern Telecommunications and Control Systems",
  chapter="127475",
  doi="10.1109/ICUMT.2016.7765379",
  howpublished="online",
  institution="EDAS Conference Services",
  year="2016",
  month="october",
  pages="320--325",
  publisher="EDAS Conference Services",
  type="conference paper"
}