Task 1.2: Transform Data and Perform Feature Engineering MLA-C01 Practice Questions (Page 2)

Questions 6–10

1. 6application · medium

   A data engineer needs to transform a large dataset (10 TB) stored in Amazon S3. The transformation involves joining two datasets, aggregating data, and writing the results back to S3 in Parquet format. The engineer wants to use a serverless approach to minimize operational overhead. Which service should the engineer use?

   • ASageMaker Data Wrangler
   • BAWS Glue ETL jobs
   • CAmazon EMR with Spark
   • DAWS Lambda functions
   Reveal Solution Discussion Select an answer first
2. 7foundation · easy

   A categorical feature has three unique values: 'red', 'green', and 'blue'. Which encoding technique creates three binary columns, one for each category, where exactly one column has a value of 1 for each row?

   • ALabel encoding
   • BBinary encoding
   • COne-hot encoding
   • DTokenization
   Reveal Solution Discussion Select an answer first
4. 8application · medium

   A data scientist is working on a dataset stored in Amazon S3. The dataset has 500,000 rows and 20 columns. The data scientist needs to explore the data visually, create a few new features, and then export the transformed dataset to a feature store for reuse. The data scientist wants to minimize the amount of custom code. Which AWS service should the data scientist use?

   • AUse Amazon EMR with Spark to process the data and then manually upload the features to SageMaker Feature Store.
   • BUse SageMaker Data Wrangler to visually explore the data, create features, and export the data to SageMaker Feature Store.
   • CUse AWS Glue ETL jobs with PySpark to transform the data and write to SageMaker Feature Store.
   • DUse AWS Glue DataBrew to visually explore the data, create features, and write the output to a CSV file in S3.
   Reveal Solution Discussion Select an answer first
5. 9application · medium

   A machine learning engineer is building a churn prediction model. The dataset contains a 'customer_region' column with 5 distinct values and a 'signup_channel' column with 3 distinct values. The engineer plans to use a logistic regression model. Which encoding strategy should the engineer use for these categorical features?

   • AApply binary encoding to both columns, creating 3 binary features.
   • BApply target encoding to both columns, replacing each category with the mean churn rate.
   • CApply one-hot encoding to both columns, creating 8 binary features.
   • DApply label encoding to both columns, assigning integers 0-4 and 0-2.
   Reveal Solution Discussion Select an answer first
6. 10foundation · easy

   A categorical feature contains ordinal values such as 'low', 'medium', and 'high'. Which encoding technique assigns integer values (e.g., 0, 1, 2) that preserve the order of the categories?

   • ATokenization
   • BBinary encoding
   • COne-hot encoding
   • DLabel encoding
   Reveal Solution Discussion Select an answer first