Time series prediction with periodic kernels

Marcin Michalak

doi:10.1007/978-3-642-20320-6_15

Time series prediction with periodic kernels

Marcin Michalak

Department of Computer Networks and Systems (RAU/RAU9)

Central Mining Institute

Research output: Contribution to journal › Article › peer-review

2 Citations (Scopus)

Abstract

This short article presents the new algorithm of time series prediction: PerKE. It implements the kernel regression for the time series directly without any data transformation. This method is based on the new type of kernel function-periodic kernel function-which two examples are also introduced in this paper. This new algorithm belongs to the group of semiparametric methods as it needs the initial step that separate the trend from the original time series.

Original language	English
Pages (from-to)	137-146
Number of pages	10
Journal	Advances in Intelligent and Soft Computing
Volume	95
Issue number	4
DOIs	https://doi.org/10.1007/978-3-642-20320-6_15
Publication status	Published - 1 May 2011

Keywords

Kernel methods
Periodic kernel functions
Regression
Semiparametric methods
Time series prediction

ASJC Scopus subject areas

General Computer Science

Access to Document

10.1007/978-3-642-20320-6_15

Cite this

@article{087dd8a6f1f44449ab39583394a233ad,

title = "Time series prediction with periodic kernels",

abstract = "This short article presents the new algorithm of time series prediction: PerKE. It implements the kernel regression for the time series directly without any data transformation. This method is based on the new type of kernel function-periodic kernel function-which two examples are also introduced in this paper. This new algorithm belongs to the group of semiparametric methods as it needs the initial step that separate the trend from the original time series.",

keywords = "Kernel methods, Periodic kernel functions, Regression, Semiparametric methods, Time series prediction",

author = "Marcin Michalak",

note = "Publisher Copyright: {\textcopyright} Springer-Verlag Berlin Heidelberg 2011.",

year = "2011",

month = may,

day = "1",

doi = "10.1007/978-3-642-20320-6\_15",

language = "English",

volume = "95",

pages = "137--146",

journal = "Advances in Intelligent and Soft Computing",

issn = "1867-5662",

publisher = "Springer Verlag",

number = "4",

}

TY - JOUR

T1 - Time series prediction with periodic kernels

AU - Michalak, Marcin

N1 - Publisher Copyright: © Springer-Verlag Berlin Heidelberg 2011.

PY - 2011/5/1

Y1 - 2011/5/1

N2 - This short article presents the new algorithm of time series prediction: PerKE. It implements the kernel regression for the time series directly without any data transformation. This method is based on the new type of kernel function-periodic kernel function-which two examples are also introduced in this paper. This new algorithm belongs to the group of semiparametric methods as it needs the initial step that separate the trend from the original time series.

AB - This short article presents the new algorithm of time series prediction: PerKE. It implements the kernel regression for the time series directly without any data transformation. This method is based on the new type of kernel function-periodic kernel function-which two examples are also introduced in this paper. This new algorithm belongs to the group of semiparametric methods as it needs the initial step that separate the trend from the original time series.

KW - Kernel methods

KW - Periodic kernel functions

KW - Regression

KW - Semiparametric methods

KW - Time series prediction

UR - https://www.scopus.com/pages/publications/84907504000

U2 - 10.1007/978-3-642-20320-6_15

DO - 10.1007/978-3-642-20320-6_15

M3 - Article

AN - SCOPUS:84907504000

SN - 1867-5662

VL - 95

SP - 137

EP - 146

JO - Advances in Intelligent and Soft Computing

JF - Advances in Intelligent and Soft Computing

IS - 4

ER -

Time series prediction with periodic kernels

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this