The use of Bayesian framework for kernel selection in vector machines classifiers

Авторы: Kropotov D., Ptashko N., Vetrov D.
Сборник: 10th Iberoamerican Congress on Pattern Recognition, CIARP 2005; Havana; Cuba;
Серия: Lecture Notes in Computer Science
Том: 3773
Год издания: 2005
Издательство: SPRINGER-VERLAG BERLIN
Местоположение издательства: HEIDELBERGER PLATZ 3, BERLIN, GERMANY,D-14197
Первая страница: 252
Последняя страница: 261
DOI: 10.1007/11578079_27
Аннотация: In the paper we propose a method based on Bayesian frame-work for selecting the best kernel function for supervised learning problem. The parameters of the kernel function are considered as model parameters and maximum evidence principle is applied for model selection. We describe a general scheme of Bayesian regularization, present model of kernel classifiers as well as our approximations for evidence estimation, and then give some results of experimental evaluation. В© Springer-Verlag Berlin Heidelberg 2005.
Добавил в систему: Кропотов Дмитрий Александрович

	ИСТИНА	Войти в систему Регистрация
	ИПМех РАН
	Главная Поиск Статистика О проекте Помощь

ИСТИНА