You work for a magazine distributor and need to build a machine learning model that predicts which customers will renew their subscriptions for the upcoming year. Using your company’s historical subscription and customer data as your training set, you created a TensorFlow model and deployed it to Vertex AI. You are now required to determine which customer attribute (e.g., age, gender, last purchase date) has the most predictive power for each prediction served by the model. What approach should you take to achieve this? | Google Professional Machine Learning Engineer Quiz