From help-octave-request at bevo dot che dot wisc dot edu Tue Feb 11 02:40:42 2003
Subject: Re: polyfit weirdness
From: Etienne Grossmann <etienne at isr dot isr dot ist dot utl dot pt>
To: Scott Lamb <slamb at slamb dot org>
Cc: help-octave at bevo dot che dot wisc dot edu,    Etienne Grossmann <etienne@isr.isr.ist.utl.pt>
Date: Tue, 11 Feb 2003 08:46:35 +0000

On Tue, Feb 11, 2003 at 02:08:37AM -0600, Scott Lamb wrote:
# I'm trying to use Octave (2.1.36, 2.1.44) for the first time. I want to 
# put some data in, make a trendline, and make a graph with the raw data, 
# trendline, trendline's formula, and correlation displayed, and attach 
# them to a LaTeX document when I'm done. (The sort of thing I could do 
# easily in Excel, but I want to learn a new open-source tool.)
# 
# So first I'm calling polyfit to get the trendline:
# 
#     l1 = 1e-9 * [ 578.45 545.88 435.87 404.71 365.26 ]
#     v1 =        [   0.66   0.79   1.40   1.58   1.92 ]
#     f1 = 299792458 ./ l1

f1 =

   5.1827e+14   5.4919e+14   6.8780e+14   7.4076e+14   8.2076e+14


     A problem with order of magnitude of f1 and floating-point
     limitations?


octave:63> f1 /= 1e14;
octave:64>      [p1,cv1] = polyfit(f1,v1,1)'
p1 =

   0.41679
  -1.49481

cv1 =

  0.66528  0.79417  1.37189  1.59260  1.92606

     Is that more like what you expected?

     Hth,
  
     Etienne


#     [p1,cv1] = polyfit(f1,v1,1)
# 
# ...and it doesn't return what I'd expect at all:
# 
#     p1 =
# 
#        1.9796e-15
#        2.8980e-30
# 
#     cv1 =
# 
#       1.0260  1.0872  1.3616  1.4664  1.6248
# 
# cv1 should hold its computed (near-identical to v1) values, right? But 
# they're not even close. Doing this in Excel gives a slope of 4.17e-15 
# and R^2 = .999. So these definitely aren't right. What am I doing 
# incorrectly?
# 
# Thanks,
# Scott Lamb
# 
# 
# 
# -------------------------------------------------------------
# Octave is freely available under the terms of the GNU GPL.
# 
# Octave's home on the web:  http://www.octave.org
# How to fund new projects:  http://www.octave.org/funding.html
# Subscription information:  http://www.octave.org/archive.html
# -------------------------------------------------------------
# 
# 
# 

-- 
Etienne Grossmann ------ http://www.isr.ist.utl.pt/~etienne



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------