A Reproducing Kernel Hilbert Space Approach to Functional Calibration of Computer Models

This article develops a frequentist solution to the functional calibration problem, where value of parameter in computer model is allowed vary with control variables physical system. The need motivated by engineering applications using constant results significant mismatch between outputs from and experiment. Reproducing kernel Hilbert spaces (RKHS) are used optimal function, defined as relationship that gives best prediction. function estimated through penalized least squares an RKHS-norm penalty data. An uncertainty quantification procedure also developed for such estimates. Theoretical guarantees proposed method provided terms prediction consistency consitency estimating function. tested both real synthetic data exhibits more robust performance than existing parametric state-of-art Bayesian method.