One-Step Residual Diagnostics Bottleneck
Background: Diagnostic checking of dynamic models with non-Gaussian errors often involves cumbersome multi-step procedures, such as running separate Ljung-Box tests on residuals and their squares. A one-step portmanteau test that captures nonlinear dependence comprehensively would simplify model validation.
Question / Future Work: Future work is needed to standardize and automate one-step residual diagnostics for non-Gaussian dynamic models to replace fragmented multi-step testing procedures. This requires validating the test’s performance across diverse model classes beyond mixed causal-noncausal autoregressive structures.
Why It Matters: Current diagnostic workflows are often prone to omission due to complexity; a reliable, single-step diagnostic tool could significantly improve model validation practices in applied time series analysis.
Metadata & Links
- created_at
- 2026-04-17T05:45:33Z
- source_papers