Methods for increase Kohonen neural network training effectiveness
Автор: Batalov A.S.
Журнал: Вестник Пермского университета. Математика. Механика. Информатика @vestnik-psu-mmi
Рубрика: Информатика. Информационные системы
Статья в выпуске: 3 (11), 2012 года.
Бесплатный доступ
The article provides the approaches to the normalization of training set. Besides the Kohonen neutral network constructive learning algorithm is suggested. Examines comparative analysis with other current Kohonen neural network learning algorithms, as a result of which the advantages and disadvantages of each algorithm are identified. In addition recommendations for increase Kohonen neural network training effectiveness are composed.
Neural networks, kohonen layer, learning algorithms, clustering
Короткий адрес: https://sciup.org/14729807
IDR: 14729807