Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!gatech!taco.cc.ncsu.edu!news-server.ncren.net!redstone.interpath.net!sas!mozart.unx.sas.com!saswss
From: saswss@hotellng.unx.sas.com (Warren Sarle)
Subject: Re: NN for Modelling Stochastic Processes
Originator: saswss@hotellng.unx.sas.com
Sender: news@unx.sas.com (Noter of Newsworthy Events)
Message-ID: <DMsD9G.5rs@unx.sas.com>
Date: Wed, 14 Feb 1996 22:03:16 GMT
X-Nntp-Posting-Host: hotellng.unx.sas.com
References:  <4fnl6q$sds@news.siemens.at>
Organization: SAS Institute Inc.
Lines: 16


In article <4fnl6q$sds@news.siemens.at>, yurt@cent.gud.siemens.co.at (Yurtsever Kutay PSE NLT) writes:
|> Has any study been done on the modelling
|> of AR, ARMA or ARIMA using NN (MLP, TDNN,
|> Recurrent, or even unsupervised learning) ?

If you have an AR, ARMA, or ARIMA process, then you should use an
AR, ARMA, or ARIMA model. There is very little point in using a more
general nonlinear model, since you will just lose statistical
efficiency.

-- 

Warren S. Sarle       SAS Institute Inc.   The opinions expressed here
saswss@unx.sas.com    SAS Campus Drive     are mine and not necessarily
(919) 677-8000        Cary, NC 27513, USA  those of SAS Institute.
