Google Professional Machine Learning Engineer
Get started today
Ultimate access to all questions.
Ultimate access to all questions.
You work at a subscription-based company that leverages machine learning models to predict customer behavior. Specifically, you have trained an ensemble of trees and neural networks to predict customer churn, which is the likelihood that customers will not renew their yearly subscription. Currently, the model's average prediction for churn rate across all customers is 15%. However, for one particular customer, the model predicts a 70% likelihood of churn. This customer has a product usage history of 30%, resides in New York City, and has been a customer since 1997. You need to provide a detailed explanation for why the model predicts a 70% churn rate for this individual as opposed to the average prediction of 15%. You plan to utilize Vertex Explainable AI to achieve this. What should you do?
Explanation:
The correct answer is B. Sampled Shapley explanations on Vertex Explainable AI provide a more sophisticated and model-agnostic method for understanding feature importance and contributions to predictions. This approach is specifically recommended for non-differentiable models, such as ensembles of trees and neural networks, which is the type of model used in the given scenario. Sampled Shapley values assign credit for the outcome to each feature and consider different permutations of the features, thus providing a comprehensive explanation of the prediction.