Choosing Features for Predictive Models

Scenario

You're training a supervised learning model and have many possible inputs, including raw fields, derived variables, and historical aggregates. You want a principled way to decide which features belong in the model.

Question

How would you choose features for a predictive model?

Problem

Scenario

Question

How would you choose features for a predictive model?

Example Dataset

size·240K customer records, 62 candidate featurestarget·Conversion within 30 daysfeatures·Profile, usage, GA4 aggregates, support, billing, engineered recency and ratio featuresmissing_data·5% to 22% missingness depending on sourceclass_balance·18% positive

Problem

Scenario

Question

How would you choose features for a predictive model?

Example Dataset

Problem

Scenario

Question

How would you choose features for a predictive model?

Interview Guides

Problem

Scenario

Question

Example Dataset

Problem

Scenario

Question

Example Dataset

Choosing Features for Predictive Models

Problem

Scenario

Question

Example Dataset

Problem

Scenario

Question

Example Dataset