A consistent model selection for orthogonal regression under component-wise shrinkage

Several authors developed a series of model selection criteria for determining the major frequency components in harmonic analysis. In this paper, we considered another direction of the extension. Specifically, we proposed a model selection criterion for an orthogonal regression estimated with a component-wise shrinkage method and proved the consistency of the proposed criterion. Through simple numerical examples, we verified the performance of the proposed criterion with the empirical component-wise shrinkage estimator. Our criterion is fully empirical and thus can be applied directly for practical uses.