A binary classification model for fraud detection is trained on a dataset where 99% of records are non-fraud and 1% are fraud. The model achieves 99% accuracy but fails to detect any fraud. What is the most appropriate technique to address this class imbalance?

Question

Accepted Answer

B. Apply SMOTE (Synthetic Minority Oversampling Technique) to the minority class. SMOTE generates synthetic examples of the minority (fraud) class by interpolating between existing minority examples, increasing its representation in training data and improving the model's ability to detect fraud.

A binary classification model for fraud detection is trained on a dataset where 99% of records are non-fraud and 1% are fraud. The model achieves 99% accuracy but fails to detect any fraud. What is the most appropriate technique to address this class imbalance?

Related Questions

Discussion