Sampling and compression of biomedical signals from point of view of telemedicine

An optimal sampling frequency is generally a compromise between the endeavour to catch all dynamical changes of a signal and not having any redundant samples. This problem has not been completely solved for crisp metric variables, and very little is known about fuzzy and qualitative variables. In this paper, one possible approach based on data compression is suggested for determining the optimal sampling frequency of ordinal and nominal variables. A similar problem is the sampling of fuzzy variables. The submitted approach can, under specific conditions, be extended to fuzzy variables too. I look at the problem initially from the point of view of quantitative metric variables. Methods for determining the optimal sampling frequency are usually based on the estimation of the spectrum of the examined signal. In this paper, an empirical method is used which does not need any knowledge of the spectrum.