You started working on a classification problem with time series data and achieved an area under the receiver operating characteristic curve (AUC ROC) value of 99% for training data after just a few…

Question

Options

• AAddress the model overfitting by using a less complex algorithm.
• BAddress data leakage by applying nested cross-validation during model training.
• CAddress data leakage by removing features highly correlated with the target value.
• DAddress the model overfitting by tuning the hyperparameters to reduce the AUC ROC value.

• A
  8% (2)
• B
  62% (16)
• C
  12% (3)
• D
  19% (5)

Explanation

Achieving 99% AUC ROC on training data after only a few experiments - especially with time series data - is a strong signal of data leakage, not simply overfitting. With time series, leakage commonly occurs when future timestamps are included in the training window, allowing the model to 'see the future.' Nested cross-validation with proper temporal ordering (walk-forward validation) prevents future data from leaking into training folds, exposing the true model performance. Option A (less complex algorithm) addresses overfitting but doesn't fix the root cause. Option C (removing correlated features) can help with leakage in some scenarios but is too narrow - removing one correlated feature doesn't prevent temporal leakage. Option D (tuning hyperparameters to reduce AUC) makes no sense; you don't tune a model to perform worse.

No community discussion yet for this question.