Your question is Learning Types for Fraud Detection. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
Explain the difference between supervised, unsupervised, and reinforcement learning in the context of fraud detection at Citi.
Describe the data each approach requires, the fraud-detection task each can support, and the main strengths and limitations of each. Explain which approach you would choose for an initial production system and how you would evaluate it. Mention why reinforcement learning is not automatically appropriate for offline fraud classification.
