MLS-C01 Question 324
Single answerYour team is tasked with building a machine learning model to predict customer churn for a subscription-based business. The dataset is structured and consists of labeled examples with features such as customer tenure, monthly spend, and support tickets. The team has limited machine learning expertise and tight deadlines for production deployment. Which approach would be most appropriate for this use case?
- A
Build a custom model using TensorFlow or PyTorch and deploy it on Amazon SageMaker.
- B
Use an Amazon SageMaker built-in algorithm such as XGBoost for the prediction task.
- C
Use Amazon SageMaker Autopilot to automatically build and deploy the best model for the dataset.
- D
Outsource the model development to a third-party vendor specializing in machine learning.
Show answer and explanation
Correct answer: B
Explanation
Amazon SageMaker built-in algorithms, such as XGBoost, are well-suited for structured datasets and common machine learning tasks like churn prediction. These algorithms are highly optimized and require minimal effort to implement, making them the most appropriate choice for teams with limited expertise and tight deadlines. Building a custom model or using Autopilot would either require more time or be unnecessary for this relatively straightforward task, while outsourcing would add cost and complexity.
- A. Incorrect.
Building a custom model from scratch using frameworks like TensorFlow or PyTorch requires significant machine learning expertise and time to develop, test, and deploy. This approach is not ideal given the team's limited expertise and tight deadlines.
- B. Correct.
Amazon SageMaker's built-in algorithms, such as XGBoost, are optimized for various machine learning tasks like classification, regression, and recommendation. They are easy to use, require minimal expertise, and can quickly deliver accurate models for structured datasets.
- C. Incorrect.
While Amazon SageMaker Autopilot is a powerful tool for automating machine learning workflows, it is more suited for scenarios where there is no clear understanding of which algorithm to use. In this case, the structured dataset and the team's familiarity with the task make XGBoost a more efficient choice.
- D. Incorrect.
Outsourcing model development to a third-party vendor may introduce delays, higher costs, and additional complexities. This option is not ideal given the team's tight deadlines and the availability of SageMaker's built-in algorithms.