Services & ResourcesWolfram ForumsMathGroup Archive

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

NonLInearFit Complaining about Confidence Levels

• To: mathgroup at smc.vnet.net
• Subject: [mg46193] NonLInearFit Complaining about Confidence Levels
• From: Ignacy Sawicki <spam at barrier.com>
• Date: Tue, 10 Feb 2004 00:06:22 -0500 (EST)
• Sender: owner-wri-mathgroup at wolfram.com

Hi,

I am trying to write a routing which does a lot of Gaussian fitting on a 
dataset. For this purpose I am calling the NonlinearRegress routing as 
follows:

bf1=NonlinearRegress[fitdata,gaussian[A,x1,xc,w],
    x1,{{A,peak[[1]]},{xc,peak[[2]]},{w,
        peak[[3]]}},{ConfidenceLevel\[Rule]0.9,Weights\[Rule]weights,
      RegressionReport\[Rule]BestFitParameters}]

and the routine complains with the following:

NonlinearRegress::bdcl: Warning: Value of option ConfidenceLevel -> {0.9} 
is not a number between 0 and 1. Setting to 0.95`.

I was not supplying a CL parameter at all at the beginning, with the same 
result. Fortunately, the whole thing runs and completes, so it doesn't 
matter for anything but aestetic reasons.

I'm new to the whole shebang. An easy explanation of how to fix this would 
be very gratefully received.

Thanks,

Iggy

• Follow-Ups:
  • Re: NonLInearFit Complaining about Confidence Levels
    • From: "Sseziwa Mukasa,,(978) 536-2359" <mukasa@jeol.com>