This module gives some important examples of expectation.
We get
Moments of a pdf by setting
in
this previous equation ,
Nth order moment
-
: 1st order moment,
= Mean value
-
: 2nd order moment,
= Mean-squared value (Power or energy)
-
: Higher order moments,
, give more detail about
.
Central moments
Central moments are moments about the
centre or
mean of a distribution,
Nth order central moment
Some important parameters from central moments of a pdf are:
- Variance,
:
- Standard deviation,
.
- Skewness,
:
if the pdf of
is
symmetric about
, and becomes more positive if the tail of the
distribution is heavier when
.
- Kurtosis (or excess),
:
for a Gaussian pdf and becomes more positive for
distributions with heavier tails.
Skewness and kurtosis are normalized by dividing the
central moments by appropriate powers of
to make them
dimensionless. Kurtosis is usually offset by
to make it zero for Gaussian
pdfs.
Example: central moments of a normal distribution
The normal (or Gaussian) pdf with zero mean is given by:
What is the
th order central
moment for the Gaussian?
Since the mean is zero, the
th
order central moment is given by
is a function of
and therefore is symmetric about zero. So all the
odd-order moments will integrate to zero (including thelst-order moment, giving zero mean). The even-order moments
are then given by:
where
is even. The integral is
calculated by substituting
to give:
Here
is the Gamma function, which is defined as an
integral for all real
and is rather like the factorial function but
generalized to allow non-integer arguments. Values of theGamma function can be found in mathematical tables. It is
defined as follows:
and has the important (factorial-like) property that
The following results hold for the Gamma function (see belowfor a way to evaluate
etc.):
and hence
Hence
- Valid pdf,
:
as required for a valid pdf.
The normalization factor
in the expression for the pdf of a unit
variance Gaussian (e.g.
) arises directly from the above result.
- Mean,
:
so the mean is zero.
- Variance,
:
Therefore standard deviation =
.
- Skewness,
:
so the skewness is zero.
- Kurtosis,
:
Hence
Evaluation of the gamma function
From the definition of
and substituting
:
Using the following squaring trick to convert this to a 2-Dintegral in polar coordinates:
and so (ignoring the negative square root):
Hence, using
:
The case for
is straightforward:
so