Task 1.3: Ensure Data Integrity and Prepare Data for Modeling MLA-C01 Practice Questions (Page 5)

Questions 21–25

1. 21application · medium

   A machine learning team is developing a model to predict loan approval. They suspect that the dataset may contain bias against a certain demographic group. The team wants to use an AWS service to detect and quantify bias before training the model. Which service should they use?

   • AAmazon SageMaker Clarify
   • BAWS Glue DataBrew
   • CAmazon SageMaker Data Wrangler
   • DAWS CloudTrail
   Reveal Solution Discussion Select an answer first
2. 22expert · hard

   A data engineer is preparing a dataset for a machine learning model that will be used by a third-party vendor. The dataset contains customer names, email addresses, and purchase history. The engineer needs to share the dataset while protecting customer privacy and ensuring that the data cannot be re-identified. Which technique should they use?

   • ATokenization
   • BPseudonymization
   • CMasking
   • DAnonymization
   Reveal Solution Discussion Select an answer first
4. 23foundation · easy

   A healthcare company is preparing patient data for a machine learning model. The dataset includes medical records that are subject to HIPAA regulations. Which type of data is being handled?

   • AProtected Health Information (PHI)
   • BPersonally Identifiable Information (PII)
   • CFinancial data
   • DNon-sensitive data
   Reveal Solution Discussion Select an answer first
5. 24foundation · easy

   A dataset contains a column with customer email addresses. Before using this data for model training, the team wants to replace the email addresses with random, non-reversible values while preserving the format for testing. Which technique is most appropriate?

   • AData augmentation
   • BEncryption
   • CTokenization
   • DNormalization
   Reveal Solution Discussion Select an answer first
6. 25application · medium

   A data scientist is preparing a dataset for a classification model. The dataset is ordered by date, and the target variable is more likely to be positive in later records. The data scientist wants to reduce prediction bias that could arise from this temporal ordering. Which data preparation step should they perform?

   • AUse the first 80% of records for training and the last 20% for testing
   • BApply data augmentation to the minority class
   • CRemove all records from the first month
   • DShuffle the dataset before splitting into training and test sets
   Reveal Solution Discussion Select an answer first