A meta-learning method to select the kernel width in Support Vector Regression

SOARES, Carlos, Pavel BRAZDIL and Petr KUBA. A meta-learning method to select the kernel width in Support Vector Regression. Machine Learning Journal. Netherlands: Kluwer Academic Publishers, 2004, vol. 54, No 3, p. 195-209. ISSN 0885-6125.

Basic information

• Original name: A meta-learning method to select the kernel width in Support Vector Regression
• Name in Czech: Metoda meta-učení pro volbu šířky jádra
• Authors: SOARES, Carlos (620 Portugal), Pavel BRAZDIL (826 United Kingdom of Great Britain and Northern Ireland) and Petr KUBA (203 Czech Republic, guarantor).
• Edition: Machine Learning Journal, Netherlands, Kluwer Academic Publishers, 2004, 0885-6125.

Other information

• Original language: English
• Type of outcome: Article in a journal
• Field of Study: 10201 Computer sciences, information science, bioinformatics
• Country of publisher: Netherlands
• Confidentiality degree: is not subject to a state or trade secret
• WWW: URL
• Impact factor: Impact factor: 3.258
• RIV identification code: RIV/00216224:14330/04:00010093
• Organization unit: Faculty of Informatics
• UT WoS: 000188925100002
• Keywords in English: meta-learning; parameter setting; support vector machines; Gaussian kernel; learning rankings
• Tags: Gaussian kernel, learning rankings, meta-learning, parameter setting, Support Vector Machines
• Tags: International impact, Reviewed

Abstract

The Support Vector Machine algorithm is sensitive to the choice of parameter settings. If these are not set correctly, the algorithm may have a substandard performance. Suggesting a good setting is thus an important problem. We propose a meta-learning methodology for this purpose and exploit information about the past performance of different settings. The methodology is applied to set the width of the Gaussian kernel. We carry out an extensive empirical evaluation, including comparisons with other methods (fixed default ranking; selection based on cross-validation and a heuristic method commonly used to set the width of the SVM kernel). We show that our methodology can select settings with low error while providing significant savings in time. Further work should be carried out to see how the methodology could be adapted to different parameter setting tasks.

Abstract (in Czech)

The Support Vector Machine algorithm is sensitive to the choice of parameter settings. If these are not set correctly, the algorithm may have a substandard performance. Suggesting a good setting is thus an important problem. We propose a meta-learning methodology for this purpose and exploit information about the past performance of different settings. The methodology is applied to set the width of the Gaussian kernel. We carry out an extensive empirical evaluation, including comparisons with other methods (fixed default ranking; selection based on cross-validation and a heuristic method commonly used to set the width of the SVM kernel). We show that our methodology can select settings with low error while providing significant savings in time. Further work should be carried out to see how the methodology could be adapted to different parameter setting tasks.

Links

• MSM 143300003, plan (intention): Name: Interakce člověka s počítačem, dialogové systémy a asistivní technologie

Investor: Ministry of Education, Youth and Sports of the CR, Human-computer interaction, dialog systems and assistive technologies