Feature Selection for Supervised Models

Scenario

You're training a supervised model and have a large set of candidate features, some of which may be redundant, noisy, or unstable across samples.

Question

How would you approach feature selection for a model?

Problem

Scenario

You're training a supervised model and have a large set of candidate features, some of which may be redundant, noisy, or unstable across samples.

Question

How would you approach feature selection for a model?

Example Dataset

Size·1.2M TikTok For You impressions, 180 candidate featuresTarget·Binary engagement labelClass balance·14% positiveFeature types·Numerical, categorical, recency, count, and cross features

Problem

Scenario

You're training a supervised model and have a large set of candidate features, some of which may be redundant, noisy, or unstable across samples.

Question

How would you approach feature selection for a model?

Example Dataset

Size·1.2M TikTok For You impressions, 180 candidate featuresTarget·Binary engagement labelClass balance·14% positiveFeature types·Numerical, categorical, recency, count, and cross features

Problem

Scenario

You're training a supervised model and have a large set of candidate features, some of which may be redundant, noisy, or unstable across samples.

Question

How would you approach feature selection for a model?

Example Dataset

Size·1.2M TikTok For You impressions, 180 candidate featuresTarget·Binary engagement labelClass balance·14% positiveFeature types·Numerical, categorical, recency, count, and cross features

Interview Guides

Problem

Scenario

Question

Example Dataset

Problem

Scenario

Question

Example Dataset

Feature Selection for Supervised Models

Problem

Scenario

Question

Example Dataset

Problem

Scenario

Question

Example Dataset