https://bugs.freedesktop.org/show_bug.cgi?id=87790

mencaraglia <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[email protected]

--- Comment #2 from mencaraglia <[email protected]> ---
Created attachment 111428
  --> https://bugs.freedesktop.org/attachment.cgi?id=111428&action=edit
spreadsheet with data (sheet base) and two fits (1 and 2 parameters)

Using LIBREOFFICE 4.3.5.2 build 3a87456aaa6a95c63eea1c1b3201acedf0751bd5 with
Kubuntu 14.04

Original data in sheet 'base'; I need to fit a power law with y data in col C,
x data in col A. Fit done using linearization log(y) log(x)

Case 1 two parameters log(y) = a * log(x) + b, sheet 2 param

The data are in range d6/e55 (logy -> col d, logx col. E) Results from function
regr.lin are in range D68/e72)   Result agree with excel, with what obtained
using functions slope(), intercept(), correlation() etc and with what obtained
using standard textbook

Case 2 one parameter log(y') = A * log(x)
y' is obtaine by simple shift ( y' = y / y0).  Data ad resultf from regr lin
same locations.  
The problem in in the R2 value D70 which is 0,9875 while I expect 0,7364 (as
fraction of explained y variance with  respect to total variance cell C70) or
0,8764 standar pearson correlation)

Excel returns 0.7364

-- 
You are receiving this mail because:
You are the assignee for the bug.
_______________________________________________
Libreoffice-bugs mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/libreoffice-bugs

Reply via email to