Experimental analysis of Support Vector Machines with different kernels based on non-intrusive monitoring data

Takashi Onoda, Hiroshi Murata, Gunnar Rätsch, Klaus Muller

Research output: Chapter in Book/Report/Conference proceedingConference contribution

15 Citations (Scopus)

Abstract

The estimation of the states of household electric appliances has served as the first application of Support Vector Machines in the power system research field [10]. Thus, it is imperative for power system research field to evaluate the Support Vector Machine on this task from a practical point of view. In this paper, we use the data proposed in [10] for this purpose. We put particular emphasis on comparing different types of Support Vector Machines obtained by choosing different kernels. We report results for polynomial kernels, radial basis function kernels, and sigmoid kernels. In handwritten digit recognition research, all results for the three different kernels achieved almost same error rates. However, in the estimation of the states of household electric appliances, the results for the three different kernels achieved different error rates. We also put particular emphasis on comparing different capacity of Support Vector Machines obtained by choosing different regularization constants and parameters of kernels. The results show that the choice of regularization constants and parameters of kernels is as important as the choice of kernel functions for real world applications.

Original languageEnglish
Title of host publicationProceedings of the International Joint Conference on Neural Networks
Pages2186-2191
Number of pages6
Volume3
Publication statusPublished - 2002 Jan 1
Externally publishedYes
Event2002 International Joint Conference on Neural Networks (IJCNN'02) - Honolulu, HI, United States
Duration: 2002 May 122002 May 17

Other

Other2002 International Joint Conference on Neural Networks (IJCNN'02)
CountryUnited States
CityHonolulu, HI
Period02/5/1202/5/17

Fingerprint

Support vector machines
Electric appliances
Monitoring
Polynomials

ASJC Scopus subject areas

  • Software

Cite this

Onoda, T., Murata, H., Rätsch, G., & Muller, K. (2002). Experimental analysis of Support Vector Machines with different kernels based on non-intrusive monitoring data. In Proceedings of the International Joint Conference on Neural Networks (Vol. 3, pp. 2186-2191)

Experimental analysis of Support Vector Machines with different kernels based on non-intrusive monitoring data. / Onoda, Takashi; Murata, Hiroshi; Rätsch, Gunnar; Muller, Klaus.

Proceedings of the International Joint Conference on Neural Networks. Vol. 3 2002. p. 2186-2191.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Onoda, T, Murata, H, Rätsch, G & Muller, K 2002, Experimental analysis of Support Vector Machines with different kernels based on non-intrusive monitoring data. in Proceedings of the International Joint Conference on Neural Networks. vol. 3, pp. 2186-2191, 2002 International Joint Conference on Neural Networks (IJCNN'02), Honolulu, HI, United States, 02/5/12.
Onoda T, Murata H, Rätsch G, Muller K. Experimental analysis of Support Vector Machines with different kernels based on non-intrusive monitoring data. In Proceedings of the International Joint Conference on Neural Networks. Vol. 3. 2002. p. 2186-2191
Onoda, Takashi ; Murata, Hiroshi ; Rätsch, Gunnar ; Muller, Klaus. / Experimental analysis of Support Vector Machines with different kernels based on non-intrusive monitoring data. Proceedings of the International Joint Conference on Neural Networks. Vol. 3 2002. pp. 2186-2191
@inproceedings{7c855f1eb192486d8d2a9807fabeb199,
title = "Experimental analysis of Support Vector Machines with different kernels based on non-intrusive monitoring data",
abstract = "The estimation of the states of household electric appliances has served as the first application of Support Vector Machines in the power system research field [10]. Thus, it is imperative for power system research field to evaluate the Support Vector Machine on this task from a practical point of view. In this paper, we use the data proposed in [10] for this purpose. We put particular emphasis on comparing different types of Support Vector Machines obtained by choosing different kernels. We report results for polynomial kernels, radial basis function kernels, and sigmoid kernels. In handwritten digit recognition research, all results for the three different kernels achieved almost same error rates. However, in the estimation of the states of household electric appliances, the results for the three different kernels achieved different error rates. We also put particular emphasis on comparing different capacity of Support Vector Machines obtained by choosing different regularization constants and parameters of kernels. The results show that the choice of regularization constants and parameters of kernels is as important as the choice of kernel functions for real world applications.",
author = "Takashi Onoda and Hiroshi Murata and Gunnar R{\"a}tsch and Klaus Muller",
year = "2002",
month = "1",
day = "1",
language = "English",
volume = "3",
pages = "2186--2191",
booktitle = "Proceedings of the International Joint Conference on Neural Networks",

}

TY - GEN

T1 - Experimental analysis of Support Vector Machines with different kernels based on non-intrusive monitoring data

AU - Onoda, Takashi

AU - Murata, Hiroshi

AU - Rätsch, Gunnar

AU - Muller, Klaus

PY - 2002/1/1

Y1 - 2002/1/1

N2 - The estimation of the states of household electric appliances has served as the first application of Support Vector Machines in the power system research field [10]. Thus, it is imperative for power system research field to evaluate the Support Vector Machine on this task from a practical point of view. In this paper, we use the data proposed in [10] for this purpose. We put particular emphasis on comparing different types of Support Vector Machines obtained by choosing different kernels. We report results for polynomial kernels, radial basis function kernels, and sigmoid kernels. In handwritten digit recognition research, all results for the three different kernels achieved almost same error rates. However, in the estimation of the states of household electric appliances, the results for the three different kernels achieved different error rates. We also put particular emphasis on comparing different capacity of Support Vector Machines obtained by choosing different regularization constants and parameters of kernels. The results show that the choice of regularization constants and parameters of kernels is as important as the choice of kernel functions for real world applications.

AB - The estimation of the states of household electric appliances has served as the first application of Support Vector Machines in the power system research field [10]. Thus, it is imperative for power system research field to evaluate the Support Vector Machine on this task from a practical point of view. In this paper, we use the data proposed in [10] for this purpose. We put particular emphasis on comparing different types of Support Vector Machines obtained by choosing different kernels. We report results for polynomial kernels, radial basis function kernels, and sigmoid kernels. In handwritten digit recognition research, all results for the three different kernels achieved almost same error rates. However, in the estimation of the states of household electric appliances, the results for the three different kernels achieved different error rates. We also put particular emphasis on comparing different capacity of Support Vector Machines obtained by choosing different regularization constants and parameters of kernels. The results show that the choice of regularization constants and parameters of kernels is as important as the choice of kernel functions for real world applications.

UR - http://www.scopus.com/inward/record.url?scp=0036085407&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036085407&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0036085407

VL - 3

SP - 2186

EP - 2191

BT - Proceedings of the International Joint Conference on Neural Networks

ER -