Accession Number : ADA289197

Title :   Evolving Cascade-Correlation Networks for Time-Series Forecasting.

Descriptive Note : Professional paper,

Corporate Author : NAVAL COMMAND CONTROL AND OCEAN SURVEILLANCE CENTER RDT AND E DIV SAN DIEGO CA

Personal Author(s) : McDonnell, J. R. ; Waagen, D. E.

PDF Url : ADA289197

Report Date : OCT 1994

Pagination or Media Count : 12

Abstract : This investigation applies evolutionary search to the cascade-correlation learning network. Evolutionary search is used to find both the input weights and input connectivity of candidate hidden units. A time-series prediction example is used to demonstrate the capabilities of the proposed approach.

Descriptors :   *FORECASTING, *TIME SERIES ANALYSIS, *MATHEMATICAL PREDICTION, MATHEMATICAL MODELS, INPUT, REPRINTS, NEURAL NETS, SEARCHING, WEIGHT, EVOLUTION(GENERAL), SYSTEMS APPROACH, LEARNING.

Subject Categories : Numerical Mathematics
      Operations Research

Distribution Statement : APPROVED FOR PUBLIC RELEASE