[Date Index] [Thread Index] [Author Index]

Re: regress versus fit -

• To: mathgroup at smc.vnet.net
• Subject: [mg61526] Re: [mg61493] regress versus fit -
• From: "David Annetts" <davidannetts at aapt.net.au>
• Date: Fri, 21 Oct 2005 00:38:11 -0400 (EDT)
• Sender: owner-wri-mathgroup at wolfram.com

Hi Chris,

> It seems like the diagnostics returned by Regress are okay, 
> but they don't have any significance arithmetic ability. 
> Looks like I will be doing each problem twice... once for the 
> regression report, and once for the "significance arithmetic" 
> available with Fit.
> David, do you work for WRI? You said you looked at the "code" 
> for Regress... I was thinking about filing a bug report or 
> RFE to have Regress' BestFit return the same result as Fit 
> for "bignum"/"sig.
> arithmetic" inputs. Good idea y/n?

Sounds like a great idea to me.  That best fit coefficients have higher
precision than the data they approximate doesn't seem right.  

I don't work for WRI.  

Looking at the code is no great mystery, at least in this case.
Statistics`LinearRegression is a standard package, which means that it
resides somewhere on the computer.  All I did was go to the directory
StringJoin[$InstallationDirectory, "\\AddOns\\StandardPackages"] and then to
the Statistics subdirectory where I could open LinearRegression.m with a
text editor (you can also use the Front End).  If the package is not
encrypted, then you can read it, and if it is, then it's for a good reason.

You can do this for standard packages which are not in Experimental,
Developer or RealTime3D contexts, and it's a great exercise for those who
want to get deeper into the system, or those who need a bit more than is
provided and are willing to get their hands dirty.

Regards,

Dave.