DocumentCode
358060
Title
Using neural nets for time series forecasting
Author
Filaretov, G.F. ; Averchenkov, E.O.
Author_Institution
Moscow Power Eng. Inst., Russia
Volume
1
fYear
1999
fDate
1999
Firstpage
249
Abstract
The purpose of the work is a comparison of statistical and neural network approaches of time series models of autoregressive moving average type construction. The basic method of research is simulation. On the basis of the results analysis, recommendations for network topology choice for neural modeling of autoregressive (AR), moving average (MA) and mixed (ARMA) processes of the various orders p, q are produced
Keywords
autoregressive processes; moving average processes; neural nets; simulation; statistical analysis; time series; autoregressive moving average type construction; mixed ARMA processes; network topology choice; neural modeling; neural nets; neural network approaches; results analysis; simulation; time series forecasting; time series models; Artificial neural networks; Electronic mail; Equations; Network topology; Neural networks; Power engineering; Predictive models; Random processes; Share prices; Time series analysis;
fLanguage
English
Publisher
ieee
Conference_Titel
Science and Technology, 1999. KORUS '99. Proceedings. The Third Russian-Korean International Symposium on
Conference_Location
Novosibirsk
Print_ISBN
0-7803-5729-9
Type
conf
DOI
10.1109/KORUS.1999.875917
Filename
875917
Link To Document