A Study on Compressed Vector-Valued Kernel Ridge Regression for the Parametric Modeling of High-Speed Interconnects

This paper investigates the performance of a compressed implementation of the vector-valued kernel ridge regression (VV-KRR) technique for the parametric modeling of the frequency response of a high-speed interconnect. The proposed approach combines data compression with efficient training to reduce computational cost while preserving model accuracy. Its performance is assessed in terms of the number of model coefficients and prediction accuracy as the number of training samples increases, demonstrating the potential of the proposed formulation for scalable and data-efficient parametric modeling.