CODEWITHSUNDEEP
machine-learningscikit-learnrandom-forest
animal
animal

Reputation: 1004

Feature importance of the predicted value in random forest?

I did my price prediction using RandomForest Regressor with Sklearn . I am able to get the feature importance using feature_importances_ function of Randomforest regressor. Now i want to know which features are affecting the individual results of the data that needs to be predicted. I cannot share the data as it is confidential just think of it as a medical claim file with 8 columns which have correlation with the actual price. Pardon me if this question sound noob i am new to machine learning world. Please give me some guidance as i am stuck up.

Upvotes: 0

Views: 582

Answers (1)

Qusai Alothman
Qusai Alothman

Reputation: 2072

What you want is called "model interpretation".

There is a famous trade-off between flexibility and interpretability. See here for a short explanation.

Now, Random Forests are fairly flexible, hence are hard to interpret. Some people even claim that this is impossible.

Keeping in mind the above, there are some people who tried to do that.
See here for a way of doing this, and here for a code example.

Upvotes: 2

Related Questions