Search papers, labs, and topics across Lattice.
This paper introduces an XMSE-aware mixed estimator that interpolates between maximum likelihood (ML) and empirical Bayes (EB) shrinkage, addressing the limitations of kernel-based EB estimation under poor parameter alignment. By deriving a closed-form oracle mixing weight, the authors ensure that the proposed estimator performs at least as well as both ML and EB estimators in terms of excess mean squared error. The methodology is validated through finite impulse response simulations and benchmarks, demonstrating its ability to retain regularization benefits while adapting to kernel misspecification.
An XMSE-aware estimator adapts between maximum likelihood and empirical Bayes, ensuring optimal performance even under kernel misspecification.
Empirical Bayes (EB) estimators can match the first-order asymptotic risk of maximum likelihood (ML) while behaving very differently at second order: recent excess mean squared error (XMSE) analysis shows that kernel-based EB estimation may be worse than ML when the kernel is poorly aligned with the true parameter. This paper turns that diagnostic into a design principle. We propose an XMSE-aware mixed estimator that interpolates between ML and EB shrinkage. Its fixed-weight XMSE is a scalar quadratic, yielding a closed-form oracle mixing weight that is no worse than both ML and the base EB estimator at the XMSE scale. A plug-in implementation based on finite-sample XMSE approximations is proved consistent, with a second-order oracle regret rate for an interior oracle weight. We further establish a transfer of the regret bound to the fixed-weight risk curve evaluated at the selected weight, a thresholded boundary rule, and extensions to compact kernel families and to finite and growing kernel dictionaries with high-probability oracle bounds. Finite impulse response simulations with SURE-tuned, hard-selection, and trace-corrected baselines, together with the public Silverbox and Cascaded Tanks benchmarks, show that the proposed estimator retains most of the benefit of regularization when it is helpful and retreats toward ML under kernel misspecification, with an identified finite-de analyzed on the benchmarks.
