### kurtosis

(measure of the probability of a distribution's tails)

The term **kurtosis** refers to the degree to which a probability
distribution has tails, regions of small probability extending
high and/or low from the highest-probability distribution.
A widely-used measure (due to Karl Pearson)
is the fourth **Pearson moment** of the distribution.
A formula:

Σ (Y_{i}-m)^{4}
kurtosis = —————————
Σ (Y_{i}-m)^{2}

- Y
_{i} - probability density.
- m - mean of all the Y
_{i}.

This measure of the **normal distribution** is always 3,
so sometimes 3 is subtracted, which is called the
**excess kurtosis**.

(*statistics,measure*)
**Further reading:**

http://en.wikipedia.org/wiki/Kurtosis

Index