random forest weight calculation method in rapidminer node

INHYEOK_SONG

I am studying feature importance. A question that arises while studying is that python sklearn random forest calculates feature importance through MDI, and I want to know how the random forest in rapidminer calculates weights.

1. MDI(Mean Decrease in Impurity) Importance

2. Permutation Importance

3. Drop Column Importance

yyhuang

Hi @INHYEOK_SONG,

To understand the attribute weights calculated for Random Forest in RapidMiner, please refer to the open sourced GitHub page.

According to the documentation, the output port for feature weight from RF model returns the

the sum of improvements the selection of a given Attribute provided at a node. The amount of improvement is dependent on the chosen criterion.

Hope it helps!

YY

INHYEOK_SONG

@yyhuang

Thanks for your reply.

 Since I only learned python, it was difficult to interpret the java code, so I left the above question.

I wanted to know which of the 1, 2, 3 items of the above question are the answers I want.

INHYEOK_SONG

@yyhuang

https://towardsdatascience.com/the-mathematics-of-decision-trees-random-forest-and-feature-importance-in-scikit-learn-and-spark-f2861df67e3

Could you please tell me if it is like the spark random forest feature importance method in this article?