1
$\begingroup$

I made a hierarchical clustering with scikit :

selected_model = AgglomerativeClustering(n_clusters=8)
hierarchical_clustering8 = selected_model.fit_predict(answers)

This classification was done on the basis of 50 features and led me to 8 clusters.

How can I proceed to determine the importance of each feature in this classification ?

My goal is to determine the most important and least important features for each cluster, and to be able to explain each cluster.

$\endgroup$
1
  • $\begingroup$ Try PCA which will give you the variance of each feature which in turn might be a good indicator of feature importance. $\endgroup$
    – spectre
    Oct 24, 2021 at 11:22

1 Answer 1

1
$\begingroup$

Since you have a estimator trained and ready. You can use the created classes and train a classification mode based on these classes. I would try a Random Forest Classifier which has a built in feature importance attribute. This attribute indicates the information gain that the features impose.

$\endgroup$

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.