A Gaussian random field (GRF) within
statistics, is a
random field In physics and mathematics, a random field is a random function over an arbitrary domain (usually a multi-dimensional space such as \mathbb^n). That is, it is a function f(x) that takes on a random value at each point x \in \mathbb^n(or some other ...
involving
Gaussian probability density functions of the variables. A one-dimensional GRF is also called a
Gaussian process. An important special case of a GRF is the
Gaussian free field
In probability theory and statistical mechanics, the Gaussian free field (GFF) is a Gaussian random field, a central model of random surfaces (random height functions). gives a mathematical survey of the Gaussian free field.
The discrete version ...
.
With regard to applications of GRFs, the initial conditions of
physical cosmology
Physical cosmology is a branch of cosmology concerned with the study of cosmological models. A cosmological model, or simply cosmology, provides a description of the largest-scale structures and dynamics of the universe and allows study of f ...
generated by
quantum mechanical fluctuations during
cosmic inflation are thought to be a GRF with a nearly
scale invariant
In physics, mathematics and statistics, scale invariance is a feature of objects or laws that do not change if scales of length, energy, or other variables, are multiplied by a common factor, and thus represent a universality.
The technical ter ...
spectrum.
Construction
One way of constructing a GRF is by assuming that the field is the sum of a large number of plane, cylindrical or spherical waves with uniformly distributed random phase. Where applicable, the
central limit theorem
In probability theory, the central limit theorem (CLT) establishes that, in many situations, when independent random variables are summed up, their properly normalized sum tends toward a normal distribution even if the original variables themsel ...
dictates that at any point, the sum of these individual plane-wave contributions will exhibit a Gaussian distribution. This type of GRF is completely described by its
power spectral density
The power spectrum S_(f) of a time series x(t) describes the distribution of power into frequency components composing that signal. According to Fourier analysis, any physical signal can be decomposed into a number of discrete frequencies, ...
, and hence, through the
Wiener–Khinchin theorem
In applied mathematics, the Wiener–Khinchin theorem or Wiener–Khintchine theorem, also known as the Wiener–Khinchin–Einstein theorem or the Khinchin–Kolmogorov theorem, states that the autocorrelation function of a wide-sense-stationary ...
, by its two-point
autocorrelation function
Autocorrelation, sometimes known as serial correlation in the discrete time case, is the correlation of a signal with a delayed copy of itself as a function of delay. Informally, it is the similarity between observations of a random variabl ...
, which is related to the power spectral density through a Fourier transformation.
Suppose ''f''(''x'') is the value of a GRF at a point ''x'' in some ''D''-dimensional space. If we make a vector of the values of ''f'' at ''N'' points, ''x''
1, ..., ''x''
''N'', in the ''D''-dimensional space, then the vector (''f''(''x''
1), ..., ''f''(''x''
''N'')) will always be distributed as a multivariate Gaussian.
References
External links
*For details on the generation of Gaussian random fields using Matlab, se
circulant embedding method for Gaussian random field
Spatial processes
{{probability-stub