Model Building - Describe Details of This Phase, Including Activities and Associated Roles DATA-SCIENCE-FOUNDATIONS Practice Questions (Page 2)

Questions 6–10

1. 6expert · hard

   A financial services company is building a fraud detection model. The data scientist has trained a deep learning model that achieves 99% accuracy on the training data but only 85% on the validation data. The business team wants to deploy the model quickly to reduce fraud losses. What should the data scientist do?

   • AAsk the data engineer to add more GPUs to speed up training
   • BDeploy the model immediately because 85% validation accuracy is acceptable
   • CInvestigate the gap between training and validation performance to address overfitting
   • DUse the training accuracy as the reported performance metric to stakeholders
   Reveal Solution Discussion Select an answer first
2. 7application · medium

   A sports analytics company is building a model to predict player performance. The data scientist has trained a model, but the model's predictions are not accurate enough. The team needs to improve the model. Which action should the data scientist take?

   • ADeploy the model and hope it improves over time
   • BAsk the data engineer to increase the storage capacity
   • CPerform feature engineering to create new relevant features
   • DReduce the number of features to simplify the model
   Reveal Solution Discussion Select an answer first
4. 8application · medium

   A healthcare analytics team is building a predictive model to identify patients at risk of readmission. The data engineer has set up a scalable compute cluster and ensured the data is accessible. The data scientist has trained several candidate models and now needs to evaluate which one generalizes best to unseen data. Which action should the data scientist take?

   • AAsk the data engineer to deploy all candidate models to production for A/B testing
   • BSplit the data into training and validation sets and compare model performance on the validation set
   • CSelect the model with the highest training accuracy without further testing
   • DUse the entire dataset to train the model to maximize accuracy
   Reveal Solution Discussion Select an answer first
5. 9expert · hard

   A telecom company is building a churn prediction model. The data scientist has trained a model, but the data engineer notices that the model uses a feature that is not available in the production data pipeline. The team needs to address this issue before deployment. What should the team do?

   • AWork together to either make the feature available in production or remove it from the model
   • BAsk the data engineer to manually add the feature to production data
   • CDeploy the model and handle missing values in production
   • DIgnore the issue because the model performs well in training
   Reveal Solution Discussion Select an answer first
6. 10application · medium

   A financial services firm is building a fraud detection model. The data scientist has trained a logistic regression model and a random forest model. To decide which model to finalize, the team needs to assess how well each model performs on data it has not seen during training. Which approach should the team use?

   • ASelect the model that trains faster
   • BUse cross-validation to evaluate both models on multiple folds of the data
   • CCompare the models' accuracy on the training data only
   • DDeploy both models and see which one performs better in production
   Reveal Solution Discussion Select an answer first