Segment regression in JASP
Hi,
I am trying to estimate whether my regression has a turning point and thought of doing a segment regression.
Anyone knows if that is possible within JASP and if yes where to find the functionality?
Thank you!
Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Quick Links
Categories
Supported by
Comments
Hi Birte,
I don't think this is possible now but it would be a good feature request for our GitHub page (preferably with a suggestion for an R package to implement; actually, the package "segmented" seems to be perfect), see https://jasp-stats.org/2018/03/29/request-feature-report-bug-jasp/
E.J.
In the meantime it might be possible for you to manually create time variables (e.g. time1 = 0 1 2 2 2 and time2 = 0 0 0 1 2 3 has two splines/one turning point)
There's a section on piecewise models in Coulombe & Selig (2015), as well as in the famous Singer & Willett (2003) book. You can extend this with an "event" dummy variable indicating when an event has occurred, add random effects with the mixed-models module in JASP, etc.
This free software will do segment regression for you.
https://www.waterlog.info/software.htm see section C "segreg"