Newsgroups: comp.ai.neural-nets,misc.invest.technical
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!nntp.sei.cmu.edu!news.psc.edu!scramble.lm.com!news.ysu.edu!news.radio.cz!newsbreeder.radio.cz!news.radio.cz!CESspool!www.nntp.primenet.com!nntp.primenet.com!news.sprintlink.net!news-peer.sprintlink.net!news-pull.sprintlink.net!news.sprintlink.net!news-dc-5.sprintlink.net!interpath!news.interpath.net!news.interpath.net!sas!newshost.unx.sas.com!saswss
From: saswss@hotellng.unx.sas.com (Warren Sarle)
Subject: Re: Optimistic and Pessimistic Views Generated by Neural Nets
Originator: saswss@hotellng.unx.sas.com
Sender: news@unx.sas.com (Noter of Newsworthy Events)
Message-ID: <E66nBD.CF@unx.sas.com>
Date: Tue, 25 Feb 1997 23:45:12 GMT
X-Nntp-Posting-Host: hotellng.unx.sas.com
References: <5ek70e$1i3@lantana.singnet.com.sg> <E60ux5.AG7@unx.sas.com> <5es9fl$beq@lantana.singnet.com.sg>
Organization: SAS Institute Inc.
Lines: 40


In article <5es9fl$beq@lantana.singnet.com.sg>, midaz tech <midaz@singnet.com.sg> writes:
|> saswss@hotellng.unx.sas.com (Warren Sarle) wrote:
|> >
|> >In article <5ek70e$1i3@lantana.singnet.com.sg>, midaz tech <midaz@singnet.com.sg> writes:
|> >|> ...
|> >|> We are currently working on a neural net system to generate optimistic 
|> >|> and pessimistic views in addition to the neural net forecasts.  A snapshot
|> >|> of our work is attached below. As illustrated, the optimistic view is 
|> >|> represented by the upward arrows while the pessimistic view by the 
|> >|> downward arrows.
|> >
|> 
|> >That's called a "prediction interval", "forecast interval", or
|> >"confidence interval for forecasted values".
|> >
|> 
|> Re the above, the "Optimisitc and Pessimistic Views" are somewhat 
|> different from "prediction interval", "forecast interval", or
|> "confidence interval for forecasted values".  The fundamental difference 
|> is that the optimistic/pessimistic projections projection above/under the 
|> forecasted values differ in magnitude, which also varies across the time 
|> scale.  Whereas the confidence interval and others are at a fixed values 
|> (plus and minus a number of standard deviations) away from the forecasted 
|> values.

There's no rule that says a confidence interval has to be symmetric.
And "forecast interval" is sufficiently vague that it doesn't have to
be associated with a precise confidence level.

I just see no point in inventing new terms when old ones already exist--
there's already been far too much of that in the neural net literature.

-- 

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.
 *** Do not send me unsolicited commercial or political email! ***

