Application of kernel produces a continuous representation
of probability density that can be used in connection with
basic concepts of traditional exclusion principles. In our case
Chauvenet’s criterion [11] may be readily applied, with an
overall risk of excluding a sound value given approximately
by 1/(2n), where n is sample size. Accordingly, given the
cumulative distribution of kernel estimate, the tail bounds
corresponding to the above defined risk may be readily
estimated.