Answers are grounded in this paper and cite their sources.
Methodological study demonstrates finite-sample inference for nonlinear switching degrees in function spaces, indicating a robust framework for structural breaks.
This paper generalizes the Switching Degree from the linear parameter space to function spaces, addressing the lack of an effect size for nonlinear regime switching. The core solution is a dual-objective design. First, taking the L² switching degree S_L² = ‖Δf‖{L²(P_X)} / ‖θ_ref‖{L²(P_X)} as the primary inferential target, we employ series estimation on known analytic bases (Fourier/polynomial bases) and rigorously derive the explicit form of the influence function, ψ_i = (c_i / C_n) [ (1/γ)(Y_i − f₁(X_i)) 1{T_i < τ} − (1/(1−γ))(Y_i − f₂(X_i)) 1{T_i ≥ τ} ], where c_i = b(X_i)ᵀ G_n⁻¹ a and a = E[b(X) Δf(X)]; the derivation does not involve the RKHS inverse operator that appears in kernel ridge regression. Under the condition that d_n grows with n and the truncation bias is o(n^(−1/2)), the influence-function debiased estimator achieves nominal coverage in simulations (95.4%–96.8%), with bias approaching zero (|bias| < 0.003) and RMSE decreasing monotonically in n. Second, the RKHS switching degree serves as a spectral-weighted descriptive measure, and the two are unified by the spectral weighting family S_α² = ‖θ_ref‖_α^(−2) Σ_k a_k² / λ_k^α (α = 0 corresponds to the L² degree, α = 1 to the RKHS degree). The linear–nonlinear decomposition theorem holds only for the RKHS switching degree: S_RKHS² = S_lin² + S_nonlin², and a general p-dimensional RKHS orthogonal projection algorithm is provided. With an unknown boundary, coverage drops to approximately 71.4%, indicating that the n-consistency boundary heuristic asymptotic framework requires post-selection correction in finite samples, which constitutes open problem SMT-T9. Under heteroskedasticity, White's robust variance raises coverage from 63.2% to 76.4% but still falls short of the nominal level, constituting open problem SMT-T8. A conditional moment switching degree is proposed as a directional framework for heteroskedasticity robustness, including a complete derivation of the first-stage error leakage expansion; its numerical validation is an open problem. Together with Tang (2026af, 2026ag, 2026ai), this paper forms part of the mathematical foundation series for structural break detection. Research Paradigm Statement: The core methodology, research direction, and final decisions were independently directed by the author. DeepSeek assisted with code implementation, data presentation, and text drafting. The author takes full academic responsibility for the final content.
No takes yet. Share an insight, caveat, or question.
Shuiping Tang (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: