Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!godot.cc.duq.edu!news.duke.edu!news.mathworks.com!newsfeed.internetmci.com!in1.uu.net!news.interpath.net!sas!newshost.unx.sas.com!saswss
From: saswss@hotellng.unx.sas.com (Warren Sarle)
Subject: Re: Why use log of values in time-series?
Originator: saswss@hotellng.unx.sas.com
Sender: news@unx.sas.com (Noter of Newsworthy Events)
Message-ID: <Dqw96x.6nF@unx.sas.com>
Date: Sat, 4 May 1996 18:40:09 GMT
X-Nntp-Posting-Host: hotellng.unx.sas.com
References:  <4me7ap$j6s@nntp.Stanford.EDU>
Organization: SAS Institute Inc.
Lines: 37


In article <4me7ap$j6s@nntp.Stanford.EDU>, mveach@cs.stanford.edu (Marshall Veach) writes:
|> I have a question re: using NNs to predict time-series data:
|> 
|> I was taking a look @ "Forecasting the Behavior of Multivariate Time
|> Series Using Neural Networks" by Chakroborty, Mehrotra, Mohan and
|> Ranka ---- in it they convert their time-series values to logarithms
|> (they don't say what base...I assume 2?) 

It doesn't matter. Changing the base just multiplies the logs by a
constant.

|> anyway, what is the advantage
|> of doing this? Is it common practice? 

Taking logarithms has nothing to do with time series, but rather with
error functions. Many commonly used error functions are functions of the
difference, abs(target-output). Suppose you are trying to predict the
price of a stock. If the price of the stock is 10 (in whatever currency
unit) and the output of the net is 5 or 15, yielding a difference of 5,
that is a huge error. If the price of the stock is 1000 and the output
of the net is 995 or 1005, yielding the same difference of 5, that is a
tiny error. You don't want the net to treat those two differences as
equally important.

By taking logarithms, you are effectively measuring errors in terms of
ratios rather than differences, since a difference between two logs
corresponds to the ratio of the original values. This has approximately
the same effect as looking at percentage differences rather than simple
differences.


-- 

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.
