Handling Imbalanced Classification Data

Scenario

You are training a binary classifier where the positive class is much rarer than the negative class. A model with high accuracy may still miss most of the cases you care about.

Question

How would you handle an imbalanced classification dataset?

Problem

Scenario

You are training a binary classifier where the positive class is much rarer than the negative class. A model with high accuracy may still miss most of the cases you care about.

Question

How would you handle an imbalanced classification dataset?

Example Dataset

size·120K rows, 38 featurestarget·Rare-event binary labelmissing_data·Low to moderate missingness in both numeric and categorical fieldsclass_balance·4.6% positive

Problem

Scenario

You are training a binary classifier where the positive class is much rarer than the negative class. A model with high accuracy may still miss most of the cases you care about.

Question

How would you handle an imbalanced classification dataset?

Example Dataset

size·120K rows, 38 featurestarget·Rare-event binary labelmissing_data·Low to moderate missingness in both numeric and categorical fieldsclass_balance·4.6% positive

Problem

Scenario

You are training a binary classifier where the positive class is much rarer than the negative class. A model with high accuracy may still miss most of the cases you care about.

Question

How would you handle an imbalanced classification dataset?

Example Dataset

size·120K rows, 38 featurestarget·Rare-event binary labelmissing_data·Low to moderate missingness in both numeric and categorical fieldsclass_balance·4.6% positive

Interview Guides

Problem

Scenario

Question

Example Dataset

Problem

Scenario

Question

Example Dataset

Handling Imbalanced Classification Data

Problem

Scenario

Question

Example Dataset

Problem

Scenario

Question

Example Dataset