Underfitting CT-AI Practice Questions (Page 4)

Questions 16–20

1. 16foundation · easy

   How can a learning curve help detect underfitting?

   • AIf the training curve is low and the validation curve is high, it indicates underfitting.
   • BIf the training and validation curves converge at a high error, it indicates underfitting.
   • CIf the training curve decreases and the validation curve increases, it indicates underfitting.
   • DIf both curves are low and close together, it indicates underfitting.
   Reveal Solution Discussion Select an answer first
2. 17application · medium

   A team is training a logistic regression model to detect fraudulent credit card transactions. The model achieves 78% accuracy on both the training and test sets, but the business requires at least 90% accuracy. The team has already verified that the data is clean and the features are relevant. Which action is most likely to improve the model's performance?

   • ACollect more training data to help the model learn the underlying patterns better.
   • BIncrease the regularization strength to prevent the model from memorizing noise.
   • CIncrease the model's complexity by using a more flexible algorithm such as a random forest or gradient boosting.
   • DReduce the number of features to simplify the model and improve generalization.
   Reveal Solution Discussion Select an answer first
4. 18foundation · easy

   Which of the following is a typical symptom of underfitting?

   • ALow accuracy on both training and test data.
   • BHigh variance in predictions across different training sets.
   • CHigh accuracy on training data but low accuracy on test data.
   • DThe model's performance improves significantly with more training data.
   Reveal Solution Discussion Select an answer first
5. 19application · medium

   A team is training a neural network for a regression task. The model is underfitting, showing high error on both training and validation sets. The team is considering several changes. Which change is most likely to reduce the underfitting?

   • AAdd L1 regularization to encourage sparsity in the weights.
   • BIncrease the dropout rate to 0.5 to prevent overfitting.
   • CAdd more hidden layers and neurons to increase the model's capacity.
   • DReduce the learning rate to make training more stable.
   Reveal Solution Discussion Select an answer first
6. 20foundation · easy

   What does high bias in a machine learning model typically indicate?

   • AThe model has too many parameters.
   • BThe model is overfitting the training data.
   • CThe model is performing well on both training and test data.
   • DThe model is underfitting the training data.
   Reveal Solution Discussion Select an answer first