The basic method uses Gaussian elimination, which
is really quite simple. Once I would have
suggested that you visit a library for a book on
numerical analysis, but nowadays I suggest a
Google search.
http://astronomy.swin.edu.au/pbourke/analysis/gausselim/
seems ok.
David J Firth wrote:
>
> In the course of doing my homework for an MBA-level managerial econ class,
> I've been spending a lot of quality time doing regressions. I use MicroTSP
> 6.6 for the Mac most of the time but use Excel 4 for the Mac occasionally.
> So, this question is not really a homework question but a tangent.
>
> I've used Tandy 100/102 and 200 machines (along with their NEC cousin) for
> 15 years. I decided to sit down and use the 200's built-in Multiplan
> spreadsheet to do a simple regression. No problem. Set up the table, do
> the calcs and summations and voila! the b and a terms for the linear form.
> I checked the results against a few worked out examples and my spreadsheet
> is fine.
>
> Not wanting to stop there, I added a second column for a second
> independent variable and based on an explanation in my handy Kachigan
> reference, ran the numbers. Doing two simple regressions for the x
> variables against a common y variable must not be the way to go. I'm
> missing something.
>
> Since the class I'm in does not require me to do regressions the "manual"
> way, this is not an exercise that I really need to do. However, I'm
> curious now and would like to see it through.
>
> Any pointers on the manual way to go a multiple linear regression the
> "old" way (before letting the PC just do it)? If I can do it a tabular way
> in a basic spreadsheet, great. I'm not above slapping together some BASIC
> code on my antique laptop either.
>
> (Oh, if you're really wondering why ... the T200 runs some 20 hours or so
> on 4 AA batteries and another 20 hours worth can easily be in my pocket.
> The T200 screen is perfectly legible outdoors on a nice day. My Mac,
> though my trusty buddy for heavy work, can't be read outdoors and runs an
> hour or so on a charge)
>
> Thanks.
>
> Dave Firth
> MBA Student
> Otterbein College
>
> --
> Dave Firth [EMAIL PROTECTED] http://www.geocities.com/djfirth/
> ----------------------------------------------------------------------
> Think Different Apple Powerbook 160
--
Bob Wheeler --- (Reply to: [EMAIL PROTECTED])
ECHIP, Inc.
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
