Fundamentals of AI and ML—II

Question 6

A financial institution wants to implement an AI solution to predict potential fraud based on transaction data. The data consists of structured fields (like transaction amounts, times, and locations) and unstructured data (like text from transaction descriptions). The institution is considering a supervised learning approach and must clearly understand the pipeline steps to ensure an effective fraud detection system.

Which steps and AWS services are most critical for building and monitoring a robust machine learning model to detect fraudulent transactions? (Select two options.)

A. Use Amazon SageMaker Data Wrangler for exploratory data analysis (EDA) and feature engineering.

B. Use Amazon Lex to deploy a chatbot for customer feedback on detected fraudulent transactions.

C. Use Amazon SageMaker Model Monitor to track model performance and detect data drift.

D. Use AWS Glue to label transaction data as fraudulent or non-fraudulent and handle data preprocessing.

E. Use Amazon Comprehend to convert transaction descriptions into labeled data for model training.