Scikit-Learn Custom Decision Tree Leaf Types

Question

Is it possible to determine the types of leaves of a decision tree when using a sci-kit?

I would like to train random forests using more complex leaves, such as leaves containing linear regressors or Gaussians. This is likely to require defining a custom sheet type and implementing a new separation criterion. Is it possible?

Thanks.

Answer 1

This can be done, but not very reasonable.

The decision tree in sklearn is written in Cython (a hybrid of C ++ and Python) and uses a predefined list of separation criteria for Cython. This makes sklearn trees very fast, but not easily customizable.

Python, Sklearn Cython. , . , , , Python ++ node. , .

, Cython, scikit-learn . , Python.

​​, M5 Weka.

, , - .