GUM2ALA – Uncertainty Propagation Algorithm for the Adaptive Linear Approximation According to the GUM

Abstract

In machine learning, many feature extraction algorithms are available. To obtain reliable features from measured data, a propagation of measurement uncertainty is necessary for these algorithms. In this contribution, the Adaptive Linear Approximation (ALA) as one feature extraction algorithm is considered, and analytical formulas are developed for an uncertainty propagation in line with the Guide to the Expression of Uncertainty in Measurement (GUM). This extends the set of uncertainty-aware feature extraction methods, which already contains the discrete Fourier and Wavelet transform.