<abstract xmlns="http://www.w3.org/1999/xhtml"><p>In the past decades, the exponential evolution of data collection for macroeconomic databases in digital format caused a huge increase in their volume. As a consequence, the automatic organization and the classification of macroeconomic data show a significant practical value. Various techniques for categorizing data are used to classify numerous macroeconomic data according to the classes they belong to. Since the manual construction of some of the classifiers is difficult and time consuming, are preferred classifiers that learn from action examples, a process which forms the supervised classification type. A variant of solving the problem of data classification is the one of using the kernel type methods. These methods represent a class of algorithms used in the automatic analysis and classification of information. Most algorithms of this section focus on solving convex optimization problems and calculating their own values. They are efficient in terms of computation time and are very stable statistically. Shaw-Taylor, J. and Cristianini, N. have demonstrated that this type of approach to data classification is robust and efficient in terms of detection of existing stable patterns in a finite array of data. Thus, in a modular manner data will be incorporated into a space where it can cause certain linear relationship.</p></abstract>

In the past decades, the exponential evolution of data collection for macroeconomic databases in digital format caused a huge increase in their volume. As a consequence, the automatic organization and the classification of macroeconomic data show a significant practical value. Various techniques for categorizing data are used to classify numerous macroeconomic data according to the classes they belong to. Since the manual construction of some of the classifiers is difficult and time consuming, are preferred classifiers that learn from action examples, a process which forms the supervised classification type. A variant of solving the problem of data classification is the one of using the kernel type methods. These methods represent a class of algorithms used in the automatic analysis and classification of information. Most algorithms of this section focus on solving convex optimization problems and calculating their own values. They are efficient in terms of computation time and are very stable statistically. Shaw-Taylor, J. and Cristianini, N. have demonstrated that this type of approach to data classification is robust and efficient in terms of detection of existing stable patterns in a finite array of data. Thus, in a modular manner data will be incorporated into a space where it can cause certain linear relationship.

Kernel Methods for Data Classification

“Spiru Haret” University, Bucharest, Romania

International conference KNOWLEDGE-BASED ORGANIZATION

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

{"article-title":"Kernel Methods for Data Classification"}

In the past decades, the exponential evolution of data collection for macroeconomic databases in digital format caused a huge increase in their volume. As a...

Kernel Methods for Data Classification

Published Online: Jul 27, 2016

Page range: 572 - 575

DOI: https://doi.org/10.1515/kbo-2016-0098

Keywords
data, macroeconomic, classifiers, kernel methods

© 2015 Raluca-Mariana Ştefan et al., published by De Gruyter Open

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

Kernel Methods for Data Classification

Published Online: Jul 27, 2016

Page range: 572 - 575

DOI: https://doi.org/10.1515/kbo-2016-0098

Keywordsdata, macroeconomic, classifiers, kernel methods

© 2015 Raluca-Mariana Ştefan et al., published by De Gruyter Open

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

Keywords
data, macroeconomic, classifiers, kernel methods