Тип публикации: статья из журнала
Год издания: 2016
Идентификатор DOI: 10.17516/1997-1397-2016-9-4-473-480
Ключевые слова: Classification, Informative variable, Optimization of the coefficient vector of the kernel fuzziness, Small training sample
Аннотация: The problem of multidimensional object classification with small training sample is considered. The following algorithms of estimating variable informativeness are considered: Ad, Del, AdDel. A new algorithm for selecting informative variables is proposed. It is based on the optimization of the coefficient vector of the kernel fuzzПоказать полностьюiness. Some modification of this algorithm is also discussed. The comparative analysis of existing methods for selecting informative variables is presented. © Siberian Federal University. All rights reserved.
Журнал: Journal of Siberian Federal University - Mathematics and Physics
Выпуск журнала: Vol. 9, Is. 4
Номера страниц: 473-480