Re: [Gnuplot-info] Fitting to weighed data points

[Sanaloht <san...@gm...>]

Damn, that is a nice reply! You're the smartest person I ever met!

Actually, the 50 tries and 30 fails are not directly represented in the
data. Only the processed pair of two values below 1 indicating percentages.
They are the datapoints to which I fit the formula.
I can get a third value. It can either be the number of tries or "number of
tries/ maximum number of tries". That is, I think, a weight.

The GnuPlot help says: " The third item (a column number or an expression),
if present, is interpreted as the standard deviation of the corresponding y
value and is used to compute a weight for the datum, 1/s**2."

I want to let GnuPlot do the weighing for me!!!! But how? 
I don't have a standard deviation of the y value, so how can I let it be
computed?
My values are totally correct, no deviation!
In which way do I have to supply the standard deviation, or, before you
flame me again for this question: How do I have to process my third value to
get a standard deviation which I can then use to compute a weight?

Well, I guess you don't answer me on this post but maybe someone else has a
answer!

-- 
View this message in context: http://www.nabble.com/Fitting-to-weighed-data-points-tp19654203p19726626.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.