Computes a statistic related to the Fourth Central Moment. Specifically, what is computed is the sum of
(x_i - xbar) ^ 4,
where the x_i are the sample observations and xbar is the sample mean.
The following recursive updating formula is used:
Let
- dev = (current obs - previous mean)
- m2 = previous value of {@link SecondMoment}
- m2 = previous value of {@link ThirdMoment}
- n = number of observations (including current obs)
Then
new value = old value - 4 * (dev/n) * m3 + 6 * (dev/n)^2 * m2 +
[n^2 - 3 * (n-1)] * dev^4 * (n-1) / n^3
Returns Double.NaN
if no data values have been added and returns 0
if there is just one value in the data set.
Note that this implementation is not synchronized. If multiple threads access an instance of this class concurrently, and at least one of the threads invokes the increment()
or clear()
method, it must be synchronized externally.