MLA-C01 Question 291
Select 3You are tasked with building a machine learning (ML) model to predict customer churn using a dataset stored in an Amazon S3 bucket. The dataset is large, contains missing values, and has both categorical and numerical features. Which of the following steps should you take to ensure the data is properly prepared for training the model?
- A
Use AWS Glue to clean the data and handle missing values.
- B
Convert categorical features to numerical representations using one-hot encoding or embeddings.
- C
Directly split the dataset into training and testing sets without addressing missing values.
- D
Use Amazon SageMaker Data Wrangler to visually analyze and preprocess the data.
- E
Enable bucket versioning on the S3 bucket to track changes in the dataset.
Show answer and explanation
Correct answers: A, B, D
Explanation
Proper data preparation is critical for building an effective ML model. In this scenario, cleaning the data, handling missing values, and preprocessing categorical features are essential steps that can be achieved using tools like AWS Glue and SageMaker Data Wrangler. Splitting the dataset without addressing missing values or relying on bucket versioning does not contribute to the data preparation process for model training.
- A. Correct.
AWS Glue is a fully managed ETL service that can be used to clean the data and handle missing values, which is an essential step for preparing the dataset for ML training.
- B. Correct.
Converting categorical features to numerical representations using techniques like one-hot encoding or embeddings is required because most ML algorithms cannot work directly with categorical data.
- C. Incorrect.
Splitting the dataset without addressing missing values will result in poor model performance as the model will encounter incomplete data during training.
- D. Correct.
Amazon SageMaker Data Wrangler allows you to visually explore, clean, and preprocess your data, making it a valuable tool for data preparation in this scenario.
- E. Incorrect.
While enabling bucket versioning is a good practice for data management, it is not directly relevant to preparing the dataset for training the ML model.