AWS Certification Exam AI Practitioner (Practice Questions) set3

問題 2 / 65

Customer churn is often a rare event with a low occurrence frequency. What is an appropriate way to handle class imbalance during training?

Balance the classes using class weighting or synthetic samples such as SMOTE.
Train using only the majority class, ignoring positive examples.
Always perform simple random undersampling.
Evaluation is based solely on accuracy.
Assume that imbalance is not a problem.

当サイトでは、ユーザー体験の向上を目的としてCookieを使用しています。サイトの利用を継続することで、Cookieの使用に同意したものとみなされます。