Anaerobic threshold and kernel estimators

Abstract

The transition threshold from aerobic to anaerobic metabolism of athletes practising endurance sports is identified with the deflection point in the graph of cardiac frequency versus velocity. A parametric statistical background was provided (Calderoni et al., 1990) by assuming a transition from linear to parabolic regression. By interpreting the threshold as a jump-point for the first derivative of a continuous regression curve, the
use of one-sided kernel estimators can be a useful explorative treatment. Here Müller’s theory of such kernels is summarized and simulations are discussed for this type of application.