Exam AWS Certified Machine Learning - Specialty topic 1 question 100 discussion

B - stratified k-fold cross-validation will enforce the class distribution in each split of the data to match the distribution in the complete training dataset.

B is the correct answer. Use Stratified k-Fold Cross-Validation for Imbalanced Classification. Stratified train/test splits is an option too. But the question is specifically asking "cross-validation" strategy.

In summary, Option B is the most appropriate strategy for handling the imbalanced dataset and ensuring reliable performance metrics for the binary classifier.

for imbalanced data. Stratified k-fold cross-validation ensures that the distribution of the target variable is the same in each fold. This is important for binary classification problems, where the target variable is imbalanced. In this case, the disease is seen in only 3% of the population. This means that if we do not use stratified k-fold cross-validation, then there is a risk that the training and validation sets will not be representative of the actual population.

B https://towardsdatascience.com/understanding-8-types-of-cross-validation-80c935a4976d

Stratified cross validation is for unbalanced data like this!

Why K=5?

Yes, B...