Parallel realization of the solution of problems of prediction and classification for cluster based on artificial neural networks
Abstract
An approach to the parallel solution of problems of classification and prediction for the apparatus of artificial neural networks in a large dimension of input and output data. Obtain the temporal characteristics of solving the problem of training neural networks with different architectures, the number of neurons and learning algorithms. We give conditions for their effectiveness in parallelizing on a computational cluster.
Problems in programming 2010; 2-3: 115-124
Full Text:
PDF (Русский)References
Хайкин С. Нейронные сети: полный курс. – М.: Вильямс, 2008. – 1104 с.
Барсегян А.А., Куприянов М.С,. Степаненко В.В, Холод И.И.. Методы и модели анализа данных: OLAP и Data Mining. – БХВ-Петербург, 2004.
Боровиков В. Statistica. Искусство анализа данных на комьютере. – 2003. – 688 с.
Круглов В.В., Борисов В.В. Искусственные нейронные сети. Теория и практика.– М.: Горячая линия –Телеком, 2002. – 382 с.
http://parallel.ru/vvv/lec7.html
http://www.ccas.ru/paral/prog/comm_msg/intro.html
http://osp.admin.tomsk.ru/os/index.htm
Refbacks
- There are currently no refbacks.







