Canonical: https://knowlpedia.com/probability/moment/

# Moment

Expected power of a random variable, used to summarize features of its distribution

For a nonnegative integer \(k\), a **moment** of order \(k\) is an [expectation](https://knowlpedia.com/probability/expectation/) of a power of a [random variable](https://knowlpedia.com/probability/random-variable/) \(X\), typically the raw moment \(\mathbb{E}[X^k]\) or the central moment \(\mathbb{E}\!\left[(X-\mathbb{E}[X])^k\right]\), whenever these expectations exist (equivalently, when \(\mathbb{E}[|X|^k]<\infty\)).

## Remarks

Moments summarize aspects of the [distribution](https://knowlpedia.com/probability/distribution-law/); for instance the second central moment is the [variance](https://knowlpedia.com/probability/variance/) and the first raw moment is the mean \(\mathbb{E}[X]\). When a [moment generating function](https://knowlpedia.com/probability/moment-generating-function/) exists in a neighborhood of zero, its derivatives recover the raw moments.

## Examples

- If \(X\sim\mathrm{Bernoulli}(p)\), then \(X^k=X\) for all integers \(k\ge 1\), so \(\mathbb{E}[X^k]=p\) and \(\operatorname{Var}(X)=p(1-p)\).
- If \(X\sim N(0,1)\), then \(\mathbb{E}[X]=0\), \(\mathbb{E}[X^2]=1\), and all odd moments are \(0\).
- If \(X\sim\mathrm{Exp}(\lambda)\) with \(\lambda>0\), then \(\mathbb{E}[X^k]=k!/\lambda^k\) for integers \(k\ge 1\).
