AWS ML Data Preparation Quiz

1. Which AWS service can be used to store and manage prepared datasets for ML training?

Amazon S3

Amazon RDS

Amazon DynamoDB

AWS Kinesis

2. Which AWS service integrates with SageMaker for automated data quality monitoring?

AWS CloudWatch

SageMaker Model Monitor

Amazon QuickSight

AWS Athena

3. In feature selection, what does the goal of reducing dimensionality accomplish?

Increases model training time

Improves model interpretability and reduces overfitting

Removes all categorical variables

Eliminates the need for data cleaning

4. Which data quality issue involves duplicate records appearing in your dataset?

Missing values

Outliers

Data duplication

Inconsistent formatting

5. What does the AWS Glue Data Catalog serve as in a data pipeline?

A machine learning model repository

A centralized metadata repository for data assets

A real-time data streaming service

A model deployment endpoint manager

6. In SageMaker Processing, what is the primary use case?

Real-time model inference

Batch data processing and transformation at scale

Model hyperparameter optimization

Deploying models to edge devices

7. Which technique is commonly used to address class imbalance in training data?

Oversampling the minority class or undersampling the majority class

Removing all minority class samples

Ignoring the imbalance during training

Using only one feature

8. What is class imbalance in a classification dataset, and why is it problematic?

It means features have different scales; it causes poor convergence

It means one class has significantly fewer samples; it biases model toward majority class

It means missing values exist; it requires deletion of rows

It means duplicate records exist; it inflates model accuracy

9. Which approach helps identify and remove highly correlated features that may cause multicollinearity?

Correlation matrix analysis

Principal Component Analysis

Both A and B

Cross-validation

10. In feature engineering, what does creating interaction features involve?

Removing highly correlated features

Combining two or more features to create new predictive variables

Normalizing all numeric features

Converting text to numerical values

11. Which AWS service is primarily used for visual data preparation and transformation in machine learning workflows?

SageMaker Data Wrangler

AWS Glue DataBrew

Amazon EMR

AWS Lambda

12. What is the purpose of train-test splitting in machine learning data preparation?

To increase dataset size

To evaluate model performance on unseen data

To remove duplicate records

To balance class distribution

13. Which statistical method is used to standardize features to have mean 0 and standard deviation 1?

Min-max scaling

Z-score standardization

Log transformation

Quantile normalization

14. In SageMaker, what does the Data Wrangler export option allow you to do?

Automatically train a model

Export data flows as a Jupyter notebook or SageMaker Processing job

Directly deploy a model to production

Generate SQL queries only

15. Which of the following is a common approach to handle outliers in a dataset?

Increase the learning rate

Use z-score or IQR method to detect and remove/cap outliers

Train multiple models without preprocessing

Duplicate the dataset

16. What is the primary benefit of using AWS Glue Crawlers in data preparation?

Automatic schema discovery and metadata cataloging

Real-time data streaming

Model training acceleration

Automated hyperparameter tuning

17. Which AWS Glue component automatically generates code for data transformation?

Glue Crawlers

Glue ETL Jobs

Glue Data Catalog

Glue Studio visual editor

18. In feature scaling, what does normalization typically rescale data to?

0 to 100

0 to 1

-1 to 1

Mean of 0 and standard deviation of 1

19. Which technique is most appropriate for converting categorical variables into numerical representations?

Normalization

One-hot encoding

Dimensionality reduction