Underfitting CT-AI Practice Questions (Page 2)

Questions 6–10

1. 6foundation · easy

   Which of the following is a common cause of underfitting in a machine learning model?

   • AUsing a model that is too complex for the dataset size.
   • BApplying strong regularization to a complex model.
   • CHaving too many features in the dataset.
   • DInsufficient training of the model.
   Reveal Solution Discussion Select an answer first
2. 7expert · hard

   A data scientist is training a random forest model. The model shows signs of underfitting: training accuracy is 70% and test accuracy is 69%. The data scientist increases the number of trees from 100 to 500, but the accuracy does not improve. What is the most likely reason this mitigation strategy failed?

   • AThe model was actually overfitting, and adding more trees made it worse.
   • BThe number of trees was already sufficient, and the model needs more training data.
   • CIncreasing the number of trees does not increase the model's capacity; it only reduces variance, which is not the problem in underfitting.
   • DThe random forest is not suitable for this type of data, and a different algorithm is needed.
   Reveal Solution Discussion Select an answer first
4. 8foundation · easy

   Which scenario is most likely to result in underfitting?

   • AUsing a linear model to capture a non-linear relationship in the data.
   • BAdding more features to a dataset with a complex model.
   • CTraining a deep neural network on a small dataset without regularization.
   • DIncreasing the number of training epochs for a complex model.
   Reveal Solution Discussion Select an answer first
5. 9foundation · medium

   In the context of machine learning, which statement best defines underfitting?

   • AThe model achieves high accuracy on both training and validation data, indicating a good fit.
   • BThe model fails to capture the underlying patterns in the training data, resulting in poor performance on both training and unseen data.
   • CThe model performs well on training data but poorly on unseen data due to excessive complexity.
   • DThe model has high variance, causing it to react too strongly to minor fluctuations in the training data.
   Reveal Solution Discussion Select an answer first
6. 10application · medium

   A data scientist is using a support vector machine (SVM) with a linear kernel to classify text documents. The model achieves 70% F1-score on both training and test sets, but the business requires 90%. The data is known to be highly non-linear. Which strategy is most appropriate to mitigate the underfitting?

   • ADecrease the regularization parameter C to make the model simpler.
   • BIncrease the regularization parameter C to reduce the penalty on misclassifications.
   • CAdd more training data to help the linear SVM learn the non-linear pattern.
   • DSwitch to an SVM with a radial basis function (RBF) kernel to capture non-linear relationships.
   Reveal Solution Discussion Select an answer first