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

Re: regress versus fit - force through zero/forceconstant term to zero

• To: mathgroup at smc.vnet.net
• Subject: [mg61505] Re: [mg61493] regress versus fit - force through zero/forceconstant term to zero
• From: "David Annetts" <davidannetts at aapt.net.au>
• Date: Thu, 20 Oct 2005 05:01:13 -0400 (EDT)
• Sender: owner-wri-mathgroup at wolfram.com

Hi Chris,
 
> If I have:
> <<Statistics`LinearRegression`
> FAASCalPts={{1,0.032},{2,0.067},{5,0.183}}
> FAASCalFit=Fit[FAASCalPts,{cCaPos},cCaPos]
> FAASCalRegression=
>   Regress[FAASCalPts,{cCaPos},cCaPos,
>     RegressionReport\[Rule]{SummaryReport,BestFit}]
> 
> I expect the answer from Fit, but not the one from Regress. 
> Regress seems to be ignoring the fact that I only want the first order 
> (and not the zero/constant order) term.
> 
> How do I obtain fit diagnostics if I am forcing a 0 for the constant 
> order?

You need IncludeConstant->False as an option for Regress ...

	FAASCalRegression =
  		Regress[FAASCalPts, {cCaPos}, cCaPos,
	    	RegressionReport -> {SummaryReport, BestFit},
IncludeConstant -> False
		]

Regards,

Dave.