Search

Results:

Vol. 16 (2013 year), No. 4

Zaporozhtsev I.F., Sereda A.-V.I.
Artificial neural networks application to temporal variability forecast of ocean surface characteristics spatial distribution

Ocean surface characteristics maps forecasting methodology has been suggested and implemented in a series of numerical experiments (with altimetry samples). Maps dataset has been considered to be a multivariate time series constructed with regularly gridded values. The methodology is based on the idea of multilayer perceptron application to the forecasting problem.

(in Russian, стр.7, fig. 3, tables. 2, ref 10, Adobe PDF, Adobe PDF 0 Kb)