Description: On a non-linear graph, when drawing the trend line, the displayed equation is absolutly wrong. The curve is right Steps to Reproduce: create a non-linear graph display the trend line and the equation Eventuellemnt, apply this equation and compare Actual Results: false equation Expected Results: good equation Reproducible: Always User Profile Reset: No Additional Info: It is possible to determine the right equation with the results of the LOGEST function I can send you a pdf document for solution
Created attachment 155615 [details] solution This document provides a solution Do not hesitate to contact me. cordially
The attached document provides a solution Do not hesitate to contact me. cordially, Olivier D
The attached document provides a solution Do not hesitate to contact me. cordially
Hi Olivier, thank you for filing. Please attach a sample .ODS Calc spreadsheet with the chart of function included. And also, please attach a screen clip with any annotation as to what is perceived as incorrect.
Comment on attachment 155615 [details] solution corrected MIME type for PDF attachement, not a patch
Created attachment 155617 [details] description Hello, in return you will find an ODS document as an example of my description. (I show the results of LOGREG by the keyboard combination ctrl/maj/enter) This file is initially created with OpenOffice, but the problem is the same with Libre Office. Cordially, Olivier D
Hi Olivier, Sorry for my ignorance but what would be the expected result ?
Hi Xisco, This concerns the formula (exponential regression) on the chart, in my opinion it is wrong. I think it's possible to calculate the coefficient of the exponential quite simply in this way: f(X) = B.exp(A.X) with A = LN(LOGREG (...; ...; 1; 1)). For coefficient 'B' I do not know just. In French it is LOGREG, I do not know if it has another name. Thank you very much
It is not possible to show a "correct" regression equation for a _line_ chart. The reason is, that a line chart treats the x-axis values as categories, which means, that they are not numbers. In the calculation of a regression line, the categories are replaced with the numbers 1,2,3,... To get a correct equation you must use an XY-chart. My suggestions since a long time is to disable the equations for line charts.
Hello Mrs. Henschel indeed, it's perfect using X;Y graphics Usually I use the particular case with a regular step, so my graphics was consistent also without X;Y Excuse me for "the alert" Best regards, Olivier D
(In reply to Regina Henschel from comment #9) > To get a correct equation you must use an XY-chart. > My suggestions since a long time is to disable the equations for line charts. (In reply to Olivier D from comment #10) > Hello Mrs. Henschel > indeed, it's perfect using X;Y graphics can this issue be closed as WORKSFORME or should it be treated as an enhancement request?
Hello, in simple user observation, 'column' type graphs have the same problem. It only works in graphical XY, which is already very good ..
Ok, let's close