Feature Selection: Filter vs Wrapper

Medium

machine-learning LinkedIn Salesforce Palantir

You have 10,000 features and need to select the most relevant ones before training. A wrapper method (like recursive feature elimination) is too slow. Which alternative approach evaluates features independently of any model?

Multiple Choice

Embedded methods like Lasso, which perform selection during model training Filter methods, which rank features using statistical measures like correlation or mutual information Forward selection, which adds features one at a time based on model performance Principal Component Analysis, which creates new features from existing ones

Correct!

Incorrect — try again next time!

Upgrade to Pro to unlock solutions and explanations

Feature Selection: Filter vs Wrapper

Multiple Choice

Related problems

Deep-dive guides