Single and multi-step ahead prediction of instant Respiration rate in incremental exercised condition with neural network

Abstract

Nowadays cardiopulmonary exercised testing (CPET) becomes popular because of its ability to interpret
cardiac and pulmonary fitness in early stage and it is applicable to normal subjects as well as patients. But in
incremental kind of exercise, level of physical stress rise with time so there is a chance of overload to the
cardiopulmonary system in later phases. So it is always preferable that some tool or facility available in the existing
system that can provide information at early. Here we have used NARNET (nonlinear autoregressive neural network) to
predict single step and multi-step ahead Instant respiration rate (RR), recorded in incremental exercise conditions. And
it is giving a satisfactory result.