DMC Troubleshooting & Problem Resolution 2

Question 1

A classification model scores 99% accuracy on training data but 68% on the test set. What is the most likely problem?

Accepted Answer

Overfitting

Answer

A large gap between training and test performance is the classic sign of overfitting.

Question 2

Which action is MOST appropriate to reduce overfitting in a decision tree?

Accepted Answer

Prune the tree or limit its depth

Answer

Pruning or limiting depth reduces model complexity and improves generalization.

Question 3

A k-means run gives different clusters each time it is executed on the same data. What is the most likely cause?

Accepted Answer

Random centroid initialization

Answer

K-means starts from random centroids and can converge to different local optima, so fixing the seed or using multiple restarts helps.

Question 4

A fraud detection model with only 1% positive cases predicts 'not fraud' for everything and reports 99% accuracy. What is the best fix to investigate first?

Accepted Answer

Address class imbalance and evaluate with precision/recall

Answer

With imbalanced classes accuracy is misleading, so resampling or class weights plus precision/recall are needed.

Question 5

Model performance drops sharply a few months after deployment even though the code has not changed. What is the most probable cause?

Accepted Answer

Concept or data drift

Answer

Changes in the underlying data distribution or relationships degrade a deployed model over time.

DMC - Data Mining Certification Practice Test

DMC - Data Mining Certification Practice Test

DMC Troubleshooting & Problem Resolution 2