Theoretical justification for bandwidth selection
Background: High-dimensional covariance estimation in time series requires data-driven bandwidth selection for long-run variance, which currently lacks a rigorous theoretical justification within many proposed frameworks.
Question / Future Work: A formal theoretical justification for the bandwidth selection procedure used in adaptive long-run variance thresholding for autocorrelated high-dimensional data is needed to ensure the consistency and optimality of the thresholding estimator.
Why It Matters: The selection of regularization and bandwidth parameters is critical for the robustness of high-dimensional sparse estimators, yet this remains a common heuristic gap in statistical literature.
Evidence: Although we do not provide a theoretical justification for this data-driven procedure, our numerical studies show that it performs reasonably well.
Metadata & Links
- created_at
- 2026-05-17T06:07:20Z