[Gnuplot-info] linear regression

[Gavin B. <gav...@ya...>]

If I understand you properly then all you want is to
plot the data and fit a straight line. This is easy in
gnuplot.

you can set your xrange and yrange to be whatever you
want - 

set xrange[0:*]
set yrange[0:*]

if you just want your axis to cross at the 0,0 point.
See help xrange.

You can also fit your data using:

fit m*x + c 'data.dat' via m,c

which will find the values of m and c that match your
data best: type 'help fit' for a description of what
this does. You can then plot it using

plot 'data.dat', m*x + c



> 
>    1. linear regression x scale, y intercept, slope
> value (r)
> 
> --__--__--
> 
> Message: 1
> To: gnu...@li...
> From:  r <re...@eu...>
> Date: Sun, 16 Oct 2005 16:43:56 +0000 (UTC)
> Subject: [Gnuplot-info] linear regression x scale, y
> intercept, slope value
> 
> Readers,
> 
> I plotted the following data into gnuplot 37:
> 
> 0.4717197379032280	0.0051347758814704
> 0.4297935152045280	0.0073454192377213
> 0.4268659918301880	0.0094278951544300
> 0.4104133185154930	0.0099098690719099
> 0.4256433835750840	0.0121491688010438
> 0.4502303629074800	0.0162362965515261
> 0.3819880241170460	0.0155606258914339
> 0.4231518138552710	0.0156275662285550
> 0.4169680192639760	0.0170985450730598
> 0.4081115179071310	0.0176955109991023
> 0.3997037161375390	0.0194298604254745
> 0.3943597830423170	0.0206064752572522
> 0.3628031878604310	0.0236138968547855
> 0.3342620844692730	0.0275198164830450
> 0.3546313302529470	0.0251857727542219
> 0.3494154104652400	0.0246491894138839
> 0.3887173220485090	0.0354124207080008
> 0.3028774653656690	0.0277222585600318
> 0.3734566008772940	0.0329406417764763
> 0.3095530446840310	0.0487083208769308
> 0.3229024733735360	0.0404098200069229
> 0.3275910144623710	0.0550704059768284
> 
> column 1 is y axis data, 2nd column for x axis. The
> resultant graph shows the
> intercept as being the first data point, 0.005,
> 0.47!
> 
> The x axis shows tics of values 5, 10, 15, 20, 25,
> when the above data is not
> within this range. Y axs shows tics in the right
> range to the data.
> 
> I want to add a trend line to the function f(x)=mx+c
> for the above data and also
> deduce values for the intercept and the slope of
> linear regression.#
> 
> Any help?
> 
> 


	
		
__________________________________ 
Yahoo! Mail - PC Magazine Editors' Choice 2005 
http://mail.yahoo.com