[Bug 123824] Exponential curve fit y=a*exp(b*x) ignores data points with zero y-values

[bugzilla]

https://issues.apache.org/ooo/show_bug.cgi?id=123824

Regina Henschel rb.hensc...@t-online.de changed:

   What|Removed |Added

 Status|UNCONFIRMED |RESOLVED
 CC||rb.hensc...@t-online.de
 Resolution|--- |WONTFIX

--- Comment #1 from Regina Henschel rb.hensc...@t-online.de ---
With the current algorithm it is not possible to use such data pairs, because
the fitting is not done on the equation y=a*exp(b*x) but on the equation
ln(y)=ln(a)+ b*x.
Excel has disabled exponential trend lines in those cases.

-- 
You are receiving this mail because:
You are the assignee for the bug.
You are watching all bug changes.

[Bug 123824] Exponential curve fit y=a*exp(b*x) ignores data points with zero y-values

[bugzilla]

https://issues.apache.org/ooo/show_bug.cgi?id=123824

--- Comment #2 from Fred Forster fors...@uw.edu ---
Quite right, and if Excel is disabled with any y=0 data points, that is the
safe things to do.  It appears Chart just ignores any such data points, does
not alert the user and merrily goes on to calculate an incorrect curve.  This
could be addressed with an alert or at least stated somewhere that the user
needs to translate his/her own data so all x and y values are positive for
valid results. Not knowing the philosophy behind OpenOffice, it is not clear if
it is supposed to produce the same answers as Excel, or do things better when
appropriate.  Anyway, I fired up Systat's SigmaPlot and got done what I needed.
 I believe it just picks initial value for coefficients and uses something like
a method of steepest ascent to minimize least square differences, but that is
just a guess.

-- 
You are receiving this mail because:
You are the assignee for the bug.
You are watching all bug changes.