Random Survival Forrest

Axel

Does someone has ever used a Random survival Forrest model in Dataiku ?

This is a definition of what is this kind of algorithm :

A Random Survival Forest ensures that individual trees are de-correlated by 1) building each tree on a different bootstrap sample of the original training data, and 2) at each node, only evaluate the split criterion for a randomly selected subset of features and thresholds.

Thanks a lot for your answer

Axel

tgb417

@Axel

Welcome to the Dataiku community. We are so glad to have you join us.

Although, I’ve never used a Random survival forrest in Dataiku.

I do note that Dataiku can be extended with sci-kit learn models.

I do note that the family of sci-kit learn models libraries does include a model for Random Survival Forrest.

If you want to try to have a go at extending Dataiku DSS you might look at these instructions

https://knowledge.dataiku.com/latest/ml-analytics/custom-models/tutorial-custom-modeling.html

This is a little bit advanced. However, there may be others here, with more experience with both Random Survival Forest and or extending Dataiku with other Sci-kit learn models who may be helpful.

Regardless let us know how you get along with this. And if you do get this working please let us know. There are some times that this kind of model may be useful to me.

Axel

Thanks a lot Tom I will look at it and let you know Merry Christmas and Happy new Year Axel

Axel

I confirm that Random Survival Forrest is part of sci-kit learn models as explained in this article so i will try to plug it in Dataiku and will ask for help if needed

Thanks again

Axel

https://scikit-survival.readthedocs.io/en/stable/user_guide/random-survival-forest.html