What remove for similar features

[Sandy4321]

May you clarify what are you doing with colinear features What you remove from let's say 3 similar features Which one?

[hppy139]

Facing the same question. @Sandy4321 Unfortunately, in your example, the code to_drop = [column for column in upper.columns if any(upper[column].abs() > correlation_threshold)](in line 201 of feature_selector.py) will remove 2 features, remaining the 1st feature in the upper triangular correlation matrix. Does it?

[Sandy4321]

any answer pls