AIF-C01 Question 38
Select 4You are an AI practitioner tasked with developing a machine learning model to predict customer churn for a subscription-based service. Which of the following steps are part of the ML development lifecycle that you should follow to ensure the model is effective and reliable?
- A
Define the business problem and the success criteria for the model
- B
Gather and preprocess the data required for training the model
- C
Deploy the model to production without testing to meet a tight deadline
- D
Evaluate the model's performance against the defined success criteria
- E
Monitor the model's predictions in production to ensure continued accuracy
Show answer and explanation
Correct answers: A, B, D, E
Explanation
The ML development lifecycle includes defining the business problem, preparing data, training and validating the model, deploying it, and monitoring its performance in production. Each step ensures the model is effective, reliable, and aligned with business objectives. Skipping steps like validation or monitoring can lead to unreliable results or model degradation over time.
- A. Correct.
Defining the business problem and success criteria is the first step of the ML development lifecycle, ensuring that the model solves the right problem effectively.
- B. Correct.
Gathering and preprocessing data is crucial to prepare the dataset, as high-quality data is essential for training an accurate model.
- C. Incorrect.
Deploying a model without testing is not part of the ML lifecycle and would lead to unreliable predictions. Testing and validation are essential before deployment.
- D. Correct.
Evaluating the model against success criteria ensures that the model meets the defined business objectives and performs as expected.
- E. Correct.
Monitoring the model post-deployment is critical to ensure that it continues to make accurate predictions over time, especially as real-world data changes.