Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.graphics.apps.gnuplot > #2428 > unrolled thread

Questions about gnuplot command fit by a newbee.

Back to article view | Back to comp.graphics.apps.gnuplot

Contents

  Questions about gnuplot command fit by a newbee. jnw@xs4all.nl - 2014-04-27 07:04 -0700
    Re: Questions about gnuplot command fit by a newbee. Hans-Bernhard Bröker <HBBroeker@t-online.de> - 2014-04-27 17:22 +0200
      Re: Questions about gnuplot command fit by a newbee. jnw@xs4all.nl - 2014-04-27 10:51 -0700

#2428 — Questions about gnuplot command fit by a newbee.

Greetings,

I'm new to gnuplot, but I've read the documentation and searched thsi newsgroup,but I still have some questions.

The first one is: Now matter what I set FIT_LIMIT to (eg. 1.0e-14), in fit.log it always says "limit for stopping : 1e-05".  The gnuplot output to stderr correctly reports "limit for stopping : 1e-14". Is this a bug?

I'm using a linear fit, as in
  f(x) = m * x + b
  fit f(x) "<file>" using 1:4:(1) via m,b
The second question is: Is there a way to include the final values of m and b in the plot?

The x-axis is a date, as in
  set xdata time
  set timefmt "%Y-%m-%d"
  set xtics format "%d-%m"
The third question is: In what unit of time is the final value of m expressed; or how do I get the value per day from the final value of m given now?

Thanks for a great program, and for your time!

[toc] | [next] | [standalone]

#2429

On 27.04.2014 16:04, jnw@xs4all.nl wrote:


> The first one is: Now matter what I set FIT_LIMIT to (eg. 1.0e-14),
> in fit.log it always says "limit for stopping : 1e-05".  The gnuplot
> output to stderr correctly reports "limit for stopping : 1e-14". Is
> this a bug?

It may have been a bug once.  I don't see such misbehaviour in the
current release version (4.6pl5), so odds are yours is just a little out
of date.

> I'm using a linear fit, as in f(x) = m * x + b fit f(x) "<file>"
> using 1:4:(1) via m,b The second question is: Is there a way to
> include the final values of m and b in the plot?

The best way of doing that depends on the version of gnuplot you have.
Current ones should support the sprintf() method (see "help sprintf").

> The x-axis is a date, as in set xdata time set timefmt "%Y-%m-%d" set
> xtics format "%d-%m"

> The third question is: In what unit of time is the final value of m
> expressed; or how do I get the value per day from the final value of
> m given now?

The unit of m is the quotient of y axis units divided by x axis units. 
In the case at hand, that's <something> per second.

[toc] | [prev] | [next] | [standalone]

#2430

I'm running what Debian calls 4.6.0-8, so it could indeed be an old bug.

Thanks for the info on sprintf and the units; I've got the correctly scaled value of variable m in the title now.

[toc] | [prev] | [standalone]

Back to top | Article view | comp.graphics.apps.gnuplot

csiph-web