|
生物物理学报 1988
SPLINE METHOD FOR NUMERICAL DIFFERENTIATION OF DIGITAL PHYSIOLOGICAL SIGNALS AND ITS ERROR ANALYSIS
|
Abstract:
The aim of this paper is to try to solve the existing problems in resent practical differential calculation of digital physiological data. A practically numerical differential method using cubic spline function accompanied with the program routine and its error bounds has been described and discussed in detail. Formulas in term of error thoery have been developed in order to analyse the transit of random error through this numerical differentiation. Based on the analysis of errors in calculation, criteria to be used to select proper sampling frequency, sampling determination and sampling-holder is put forward. A method of resolution of extreme value of derivative curve is also introduced here. Then, illustrative examples, including the differentialti-ons of phase 0 of myocardial action potential, of thoracic impedance and of left ventricle presure, are given.