1. Science
  2. Publications
  3. Information Processing Systems
  4. 7(74)'2008
  5. Noposition encoding in tasks of neural networks prediction of time series

Noposition encoding in tasks of neural networks prediction of time series

E.V. Bodyansky, O.S. Udovenko
Annotations languages:

Is considered one of the algorithms of evaluation and determinations of acceptability of risk level when ensuring information safety. The method of the module information encoding is examined in forecasting models which will be realized on artificial neural networks. Advantages of offered approach consist of simplification of structure of neural networks, diminishing of data and possibility of the use of parallel computable procedurest, that allows to promote general efficiency of the system of prediction.
Keywords: modular code, artificial neuron network, system of remaining classes, synaptic matrix