Developing adaptive models and estimating the uncertainty of presampled spectral data

The use of model-based parameter estimation (MBPE) to develop adaptive, windowed, rational-function fitting models for presampled electromagnetic data, and for estimating the uncertainty of such data has been outlined and illustrated. For the former application, MBPE is advantageous in permitting a reduced-order, analytically continuous approximation to the discretely sampled data to be realized. For the latter, the MBPE procedure permits the uncertainty of the presampled data to be estimated in a manner analogous to linear regression when data is described by a straight line.