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

--- Comment #4 from mencaraglia <[email protected]> ---
Created attachment 111516
  --> https://bugs.freedesktop.org/attachment.cgi?id=111516&action=edit
short sample of spradsheet

I am resending the spradsheet as a short version: Data are in col A and B; same
data (but for a constant shift in y) are in col F and G.  Input of these data
is TEP 1

Below each series (A & B ; F & G) the linear best fit using  A = a B + b (col A
& B)  or  F = a' G ( col F & G); choice on how to fit from the third parameter
of regr.lin

regr.lin(A;B; 1 ; 1)
regr.lin(F;G; 0 ; 1)

Evaluation of regr-lin  is STEP 2

As a check for col A & B also evaluation using slope(A;B) intercept(A;B)
correlation(A;B) RQ(A:B) have been made

These checks are STEP 3

I have shown corresponding results using background color

Now, passing to col F and G the result from regr.lin(...) are ok for the slope
(not shown here the check);   correlation(F;G) and RQ(F:G) correspond with what
evaluated in A & B (as expected, these values are invariant) 

What I do not understand is the term on cell F56 (cell background in red)which
- as far as I understand should be a correlation - but is neither correlation()
nor RQ() and - as I to before is not equal to the value suggesteb by excel.

To be sure that what I find is clear I have also copied the results with 'paste
special - only numbers - in the area with gray background.

-- 
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