The main goal of the experimental work is to see how
well the algorithm performs in handling artifacts from IBI
signal while calculating the HRV features. For evaluation
ten visually inspected ‘No artifacts’ and three ‘With
artifacts’ measurements as discussed in the ‘Data
collection and HRV features’ section are considered. Now,
from these three ‘With artifacts’ measurements 30
corrupted measurements are generated by introducing three
random sets of artifacts and randomly placing them with
each ‘No artifacts’ measurements. Then the time and
frequency domain features are calculated for these 30
‘With artifacts’ measurements. The intention is to see how
the features values are deviated for the ‘With artifacts’
measurements compare to ‘No artifacts’ measurements
using the algorithm. The performance of the proposed
algorithm is investigated both in detecting and
interpolating artifacts. Another algorithm i.e., the
hierarchical algorithm is introduced for detecting the
outlier and results are compared with the proposed
algorithm. Third party software [19] which has applied
cubic spline interpolation is considered to evaluate the
performance of the proposed algorithm.