Thompson, A (2025) Analytical results for combined data and model uncertainty for machine learning regression. Measurement: Sensors, 38. 101788

Full text not available from this repository.

Abstract

We present a variance decomposition framework which combines data and model uncertainty for machine learning (ML) regression models. Within this framework, we present analytical expressions for the mean and variance of the output of some popular ML regression models, including linear regression and Gaussian Processes. The results extend work by the same author on trained (fixed) models recently presented in [2]. We illustrate the results in the context of state-of-health modelling of lithium-ion cells based on Electric Impedance Spectroscopy (EIS) measurements.

Item Type:	Article
Subjects:	Mathematics and Scientific Computing > Measurement Uncertainties
Divisions:	Data Science
Identification number/DOI:	10.1016/j.measen.2024.101788
Last Modified:	06 Oct 2025 09:54
URI:	https://eprintspublications.npl.co.uk/id/eprint/10163