Regression analysis treats missing values as zeros

Bug #331756 reported by Ola Hodne Titlestad on 2009-02-19
Affects Status Importance Assigned to Milestone
Lars Helge Øverland

Bug Description

When generating regression lines for report tables the calculated regression values are wrong because empty values are treated as zeros, which gives a very low line. Regression lines should be calculated based on registered values only, that is one of the major reasons for having regression in the first place...

Bob Pond pointed this out in his e-mail on 19.02.2009 commenting the SL dmeo application:

"The chart I see showing "immunization trends Kono" is a promising
development. However, it appears that the trend line is drawn by
assuming that missing values (for the most recent months) are zero
values. Can this be corrected (by using linear regression of
non-missing values)?

Changed in dhis2:
assignee: nobody → larshelge
importance: Undecided → High
milestone: none → 2.0.1final
Changed in dhis2:
status: New → Fix Committed
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers