Mean Square Prediction Error of Misspecified Gaussian Process Models

by   Thomas Beckers, et al.

Nonparametric modeling approaches show very promising results in the area of system identification and control. A naturally provided model confidence is highly relevant for system-theoretical considerations to provide guarantees for application scenarios. Gaussian process regression represents one approach which provides such an indicator for the model confidence. However, this measure is only valid if the covariance function and its hyperparameters fit the underlying data generating process. In this paper, we derive an upper bound for the mean square prediction error of misspecified Gaussian process models based on a pseudo-concave optimization problem. We present application scenarios and a simulation to compare the derived upper bound with the true mean square error.


page 1

page 2

page 3

page 4


Prediction performance after learning in Gaussian process regression

This paper considers the quantification of the prediction performance in...

Clustered Gaussian process model with an application to solar irradiance emulation

A Gaussian process has been one of the important approaches for emulatin...

A deep learning approach for analyzing the composition of chemometric data

We propose novel deep learning based chemometric data analysis technique...

Co-clustering separately exchangeable network data

This article establishes the performance of stochastic blockmodels in ad...

Efficiency Assessment of Approximated Spatial Predictions for Large Datasets

Due to the well-known computational showstopper of the exact Maximum Lik...

Direct loss minimization for sparse Gaussian processes

The Gaussian process (GP) is an attractive Bayesian model for machine le...

Please sign up or login with your details

Forgot password? Click here to reset