I would like to make a regression tree like the one in the picture. The tree was done in Cubist but I don't have that program. I do use R and Python. It seems to differ from the R packages rpart or tree in that the end nodes are linear formulas rather than just the average value. Is there any way I can do this using R or some other free software?
In the picture, NDVI, B1,B2, etc are variables. The image is from this website.
The overview of the R implementation of Cubist can be found here.
From that overview, the first part "of the algorithm is consistent with the 'M5' or Model Tree approach."
Specifically, the differences are that:
Cubist
is an R port of the Cubist GPL C code released by RuleQuest at http://rulequest.com/cubist-info.html.Using the example from
help('cubist')
and the original package announcement