Data analysis in dielectric time-domain spectroscopy

An approach for overcoming accuracy problems connected with dielectric time-domain spectroscopy is described. It is pointed out that, for an accurate analysis of data with a large dynamic range and recorded with changes in instrumental parameters such as gain switching, several steps are essential. These include (1) noise reduction, (2) adjustment for instrumental changes, called data patching, and (3) transform methods such as the Laplace transform, to bring the data into the frequency domain. The authors attempt to show how this approach can overcome accuracy problems. They explain why the common practice of interpolation on data points, e.g., with cubic or B splines, may be the worst method of noise reduction for this type of instrumental data and should be replaced by fitting the data to given functions, determined by the application