This paper introduces a theoretical framework for kernel ridge regression that accounts for varying scale parameters, which are crucial for modeling data correlation. The research establishes a condition for training-point risk consistency on uniform tensor grids with product exponential kernels: the effective local sample size must diverge to infinity. This criterion unifies existing theories and includes a deterministic regularization penalty construction to ensure consistency, with numerical experiments validating the findings across different regimes. AI
IMPACT Provides a theoretical advancement in kernel methods, potentially improving model robustness in scenarios with changing data distributions.
RANK_REASON This is a theoretical research paper published on arXiv. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →