AI-900 Question 76
Select 3A retail company wants to predict customer churn using Azure Machine Learning. They have historical customer data, including demographics, purchase history, and customer support interactions. Which of the following key steps should they include in their machine learning workflow on Azure?
- A
Clean and preprocess the customer data to address missing or inconsistent values.
- B
Deploy the model to an Azure virtual machine without testing its accuracy.
- C
Split the historical data into training and testing datasets for model evaluation.
- D
Use an unsupervised learning algorithm to predict customer churn.
- E
Train a supervised learning model using labeled data with churn outcomes.
Show answer and explanation
Correct answers: A, C, E
Explanation
To predict customer churn using Azure Machine Learning, the company must follow a systematic workflow. This includes cleaning and preprocessing the data to ensure quality, splitting the data into training and testing datasets for accurate evaluation, and training a supervised learning model since churn prediction is a supervised task. Testing the model's accuracy is also critical before deployment, and unsupervised learning is not suitable for this problem as it lacks labeled outcomes.
- A. Correct.
Correct. Cleaning and preprocessing data is a fundamental step in any machine learning workflow to ensure the quality of the input data.
- B. Incorrect.
Incorrect. Deploying a model without testing its accuracy is not recommended because it may lead to unreliable predictions.
- C. Correct.
Correct. Splitting the data into training and testing datasets is essential for evaluating the model's performance and avoiding overfitting.
- D. Incorrect.
Incorrect. Unsupervised learning is not suitable for predicting customer churn because churn prediction typically requires labeled data and is a supervised learning task.
- E. Correct.
Correct. Training a supervised learning model with labeled data (e.g., churn outcomes) is the appropriate approach for this problem.