Gaussian Process-Aided Function Comparison Using Noisy Scattered Data

This work proposes a nonparametric method to compare the underlying mean functions given two noisy datasets. The motivation for stems from an application of comparing wind turbine power curves. Comparing data presents new problems, namely need identify regions difference in input space and quantify extent that is statistically significant. Our proposed method, referred as funGP, estimates different samples using Gaussian process models. We build confidence band probability law estimated function differences under null hypothesis. Then, used hypothesis test well identifying difference. identification distinct feature, existing methods tend conduct overall stating whether are different. Understanding can lead further practical insights help devise better control maintenance strategies turbines. merit funGP demonstrated by three simulation studies four real