The interpolation in the proposed algorithm is based on
K-NN algorithm and considers data points that lie just
before or after the corrupted area. The experimental work
(Table 3) considering k=1, 2, and 3 explains that it works
better when we consider k=2 during interpolation. A third
party software Kubios HRV is also introduced to
investigate interpolation performance. It illustrates that the
proposed algorithm performs well in some cases however,
in other cases the software works better. However, in
Kubios HRV software a manual intervention is needed
while choosing the interpolation level for a signal.
Another, experimental work discussed in Table 5 for a
specific signal shows that after interpolation the new data
points explain a most of the original data points (without
artifacts) using the algorithm. So, one of the major benefit
of the proposed algorithm is that it does not change or
work on the whole signal rather it treated only the part of
the signal that is corrupted. Thereby it reduces the
possibility of changing the original aritfact free data points.