0
$\begingroup$

I need to build a Multivariate Regression Tree. Looking at Scikit-learn's Multi-output Decision Tree Regressor it seems that what they do is define as many regressor as there are dependant variables. However, as Henrik Linusson says in Multi-Output Random Forests, this is a naive approach. Is seems like what he says about the latest approach seems to be the following (referring to Random Forests):

Segal and Xiao (2011) propose an extension to the Random Forest ensemble predictor that handles multivariate responses — extension is done simply by replacing the typical univariate trees in the Random Forest with the same type of multivariate trees as those proposed by Segal (1992)

Is there any code available that does something similar, or would I have to implement it from scratch? If so, what are some good frameworks that I could use?

Thanks in advance for any help!

$\endgroup$

1 Answer 1

0
$\begingroup$

Scikit learn's documentation suggests that it is not a multiple univariate approach. Please look at this documentation.

1.10.3. Multi-output problems

https://scikit-learn.org/stable/modules/tree.html#multi-output-problems

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.