Hands-On Exercise: Identify Training and Test Data and Create an ML Model CT-AI Practice Questions (Page 3)

Questions 11–15

1. 11foundation · easy

   Which statement correctly describes the relationship between training and test datasets in a typical ML workflow?

   • AThe test set is used to train the model, and the training set is used to evaluate it.
   • BThe test set is used to tune hyperparameters, and the training set is used for final evaluation.
   • CThe training set is used to fit the model, and the test set is used to assess its performance.
   • DThe training set and test set are identical to ensure consistency.
   Reveal Solution Discussion Select an answer first
2. 12application · medium

   A team is building a model to classify customer reviews as positive or negative. They have a dataset of 10,000 reviews. They split the data into 80% training and 20% test sets. They train a support vector machine (SVM) and get 92% accuracy on the test set. Which of the following is the most important next step before deploying the model?

   • ARetrain the model on the entire dataset to maximize performance before deployment.
   • BIncrease the size of the test set to improve the accuracy estimate.
   • CCheck the model's performance on different subsets of the test set, such as reviews from different product categories.
   • DEvaluate the model on a separate validation set that was not used during training.
   Reveal Solution Discussion Select an answer first
4. 13foundation · easy

   In supervised machine learning, what is the primary purpose of the test dataset?

   • ATo augment the training data with additional samples
   • BTo train the model by adjusting its parameters
   • CTo evaluate the model's performance on unseen data
   • DTo select the best features during model training
   Reveal Solution Discussion Select an answer first
5. 14application · easy

   A data analyst is building a model to predict whether a customer will buy a product. The dataset has 20 features, including age, income, and browsing history. She trains a decision tree and gets 85% accuracy on the test set. She then trains a random forest on the same data and gets 88% accuracy on the test set. Which of the following is the most appropriate conclusion?

   • AThe random forest performs better on the test set and should be preferred.
   • BThe decision tree is better because it is simpler and easier to interpret.
   • CThe random forest is overfitting because it is more complex than the decision tree.
   • DThe models are equivalent because the accuracy difference is too small to matter.
   Reveal Solution Discussion Select an answer first
6. 15expert · hard

   A machine learning team is developing a model to diagnose a rare disease. The disease occurs in 0.1% of the population. They have a dataset of 100,000 patients, with 100 positive cases. They train a model and evaluate it on a test set with the same disease prevalence. The model achieves 99.9% accuracy. Which of the following is the most appropriate next step for the team?

   • AExamine precision, recall, and the ROC curve to understand the model's ability to detect the rare disease.
   • BCollect more data on the majority class to improve the model.
   • CDeploy the model immediately because 99.9% accuracy is excellent.
   • DUse a larger test set to get a more accurate accuracy estimate.
   Reveal Solution Discussion Select an answer first