A dataset contains text reviews labeled as either 'positive' or 'negative'. For a protected attribute 'region', the proportion of 'positive' labels is 0.6 for region A and 0.4 for region B. Which pre-training bias metric is most directly calculated from these two proportions?

Difference in proportions of labels (DPL)

A company is building a machine learning model using health records that are classified as protected health information (PHI) under HIPAA. The company plans to use AWS for data storage and processing. Which action is required to comply with HIPAA?

Store all PHI in a specific AWS region

Sign a Business Associate Agreement (BAA) with AWS

Use Amazon S3 with versioning enabled

A dataset for image classification has 10,000 images of cats and only 500 images of dogs. The team wants to balance the dataset without discarding any original images. Which strategy directly addresses this class imbalance?

Changing the loss function to ignore the minority class

Undersampling the majority class by removing 9,500 cat images

Applying random rotations and flips to all cat images

Oversampling the minority class by duplicating dog images

A company is using Amazon S3 to store sensitive customer data. They want to ensure that data is encrypted at rest using a key that they manage themselves, but they also want to rotate the key automatically. Which AWS service should they use for key management?

AWS Key Management Service (KMS)

AWSCertified Machine Learning Engineer - Associate

Domain 1Objective 3

Task 1.3: Ensure Data Integrity and Prepare Data for Modeling MLA-C01 Practice Questions (Page 2)

Part of the Content Domain 1: Data Preparation for Machine Learning (ML) domain, which accounts for 28% of the MLA-C01 exam. AWS does not publish an official question count, but from its 130-minute exam (~50–85 total, ~14–24 in this domain), expect 5–8 from this objective — we provide 28 practice questions to prepare you well beyond it. (estimate)

28questions here

6free pages

9concepts

28%of the exam

Questions 6–10

6foundation · easy
A dataset contains text reviews labeled as either 'positive' or 'negative'. For a protected attribute 'region', the proportion of 'positive' labels is 0.6 for region A and 0.4 for region B. Which pre-training bias metric is most directly calculated from these two proportions?
Select an answer first
7application · medium
A company is building a machine learning model using health records that are classified as protected health information (PHI) under HIPAA. The company plans to use AWS for data storage and processing. Which action is required to comply with HIPAA?
Select an answer first
8foundation · easy
A dataset for image classification has 10,000 images of cats and only 500 images of dogs. The team wants to balance the dataset without discarding any original images. Which strategy directly addresses this class imbalance?
Select an answer first
9application · medium
A company is using Amazon S3 to store sensitive customer data. They want to ensure that data is encrypted at rest using a key that they manage themselves, but they also want to rotate the key automatically. Which AWS service should they use for key management?
Select an answer first
10foundation · easy
A team needs a high-performance, POSIX-compliant file system for a training workload that requires very low latency and high throughput, such as large-scale deep learning. Which AWS service is optimized for this performance requirement?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to MLA-C01

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by AWS. “MLA-C01” is a trademark of its owner, used for identification only.