Article View: comp.apps.spreadsheets
Article #205Re: Excel 4.0 features?
From: Jess M Holle
Date: Fri, 16 Oct 1992 04:43
Date: Fri, 16 Oct 1992 04:43
18 lines
1428 bytes
1428 bytes
In article <1992Oct15.191726.22860@klaava.Helsinki.FI> pemakine@klaava.Helsinki.FI (Pekka Makinen) writes: > In addition to LINEST() function Excel has also a LOGEST() function, > which will give regression values for an exponential function. That > is at least some kind of a non-linear regeression. Perhaps in one sense of the term "non-liner," however they certainly use the same algorithm as for y=ax+b, except that they fit to ln(y) = ln(b) + x ln(a) which can be rewritten as y' = a'x + b'. Therefore, this is really just a special case of linear regression. What I would consider non-linear regression would be something like fitting: y = a sin(b x + c) I know of no way to linearize this. This requires a different and more complicated approach then linear regression. Jess Holle
Technical Details
Message-ID:
<1992Oct16.044302.29726@gn.ecn.purdue.edu>Path:
arc1-ca.pugleaf.net!sparky!uunet!pipex!unipalm!uknet!stl!bnrgate!nrcnet0!cunews!torn!cs.utexas.edu!zaphod.mps.ohio-state.edu!uwm.edu!linac!att!news.cs.indiana.edu!noose.ecn.purdue.edu!gn.ecn.purdue.edu!jessReferences:
<3360@ariel.its.unimelb.EDU.AU> <james.8.719150890@library.nrl.navy.mil> <1992Oct15.191726.22860@klaava.Helsinki.FI>Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!pipex!unipalm!uknet!stl!bnrgate!nrcnet0!cunews!torn!cs.utexas.edu!zaphod.mps.ohio-state.edu!uwm.edu!linac!att!news.cs.indiana.edu!noose.ecn.purdue.edu!gn.ecn.purdue.edu!jess From: jess@gn.ecn.purdue.edu (Jess M Holle) Newsgroups: comp.apps.spreadsheets Subject: Re: Excel 4.0 features? Message-ID: <1992Oct16.044302.29726@gn.ecn.purdue.edu> Date: 16 Oct 92 04:43:02 GMT References: <3360@ariel.its.unimelb.EDU.AU> <james.8.719150890@library.nrl.navy.mil> <1992Oct15.191726.22860@klaava.Helsinki.FI> Organization: Purdue University Engineering Computer Network Lines: 18