Câu 18: RAI: Risk and Artificial Intelligence

An analyst uses the K-nearest Neighbors (KNN) model to classify hedge funds using volatility and maximum drawdown after standardization. The Euclidean distance is used as the distance metric, and K is set to 5. After adding assets under management (AUM) in USD millions to the model and retraining it, the results change substantially and no longer align with historical risk assessments. Which of the following is most likely the cause?

1. A. The model should be switched from classification to regression, since the additional feature is continuous.
2. B. Euclidean distance is not an appropriate metric for this case; Manhattan distance should be used instead.
3. C. The number of nearest neighbors (K) is not adjusted after including the additional feature.
4. D. The unsealed third feature (AUM) dominates the distance metric and alters nearest neighbor relationships. — đáp án hiện tại