Xref: utzoo comp.os.os2.apps:58 comp.os.msdos.apps:745 comp.windows.ms:7868
Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!usc!samsung!munnari.oz.au!sirius.ucs.adelaide.edu.au!levels!marwk
From: marwk@levels.sait.edu.au
Newsgroups: comp.os.os2.apps,comp.os.msdos.apps,comp.windows.ms
Subject: Re: Excel:1. Regression? 2.Memory? 3. DesqView?
Message-ID: <15752.2764dfc0@levels.sait.edu.au>
Date: 11 Dec 90 13:31:44 GMT
References: <7973.275c1a76@uwovax.uwo.ca> <1990Dec8.171008.5429@herald.usask.ca>
Organization: Sth Australian Inst of Technology
Lines: 26

In article <1990Dec8.171008.5429@herald.usask.ca>, reeves@dvinci (Malcolm Reeves) writes:
> From article <7973.275c1a76@uwovax.uwo.ca>, by jansson@uwovax.uwo.ca:
>> I am trying to estimate a regression equation using Excel.  According to
>> the manual I should use the function LINEST and given the data arrays
>> Excel should compute the slopes and the intercept for me.  The problem is
>> that the only output I get is one of the slopes.  I would like to get all
>> the information for the equation.  How do I go about getting it?
>
> You've just discovered the problems associated with using "canned" software.
> It's easy to compute a least-squares regression using the arithmetic in a
> spreadsheet (I don't know excel but I can do it in Lotus, Supercalc, Quattro
> etc) then you can recover the full equation and standard errors of estimate
> for the coefficients. Another problem is that spreadsheets do not check if
> the data is suitable for linear least squares regression nor do they provide
> other norms or non-parametric alternatives when the point distribution is
> obviously not "normal".
>
> This may not help much but the spreadsheet arithmetic for least-squares
> regression is easy and you can have any information you need if you do it
> yourself :-)
>
>
You must highlight 2 adjacent cell and place the LINEST function in the left
cell - simple as that.

Ray