1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
|
.. _kriging:
Kriging
-------
Kriging (also known as Gaussian process regression) is a Bayesian
technique that aim at approximating functions (most often in order to
surrogate it because it is expensive to evaluate). In the following it
is assumed we aim at creating a surrogate model of a scalar-valued model
:math:`\cM: \vect{x} \mapsto y`. Note the implementation of
Kriging deals with vector-valued functions
(:math:`\cM: \vect{x} \mapsto \vect{y}`), without simply looping over
each output. It is also assumed the model is obtained over a design of
experiments in order to produce a set of observations gathered in the
following dataset:
:math:`\left(\left(\vect{x}^{(i)}, y^{(i)}\right), i = 1, \ldots, m\right)`.
Ultimately Kriging aims at producing a predictor (also known as a
response surface or metamodel) denoted as :math:`\tilde{\cM}`.
We put the following Gaussian process prior on the model :math:`\cM`:
.. math:: Y(\vect{x}) = \Tr{\vect{f}(\vect{x})} \vect{\beta} + Z(\vect{x})
where:
- :math:`\Tr{\vect{f}(\vect{x})} \vect{\beta}` is a general linear
model based upon a functional basis
:math:`\vect{f} = \left(f_j, j = 1, \ldots, p\right)` and a vector of
coefficients
:math:`\vect{\beta} = \left(\beta_j, j = 1, \ldots, p\right)`,
- :math:`Z` is a zero-mean stationary Gaussian process whose covariance
function reads:
.. math:: \mathbb{E}[Z(\vect{x})\,Z(\vect{x'})] = \sigma^2 R(\vect{x} - \vect{x'}, \vect{\theta})
where :math:`\sigma^2 > 0` is the variance and :math:`R` is the
correlation function that solely depends on the Manhattan distance
between input points :math:`\vect{x} - \vect{x'}` and a vector of
parameters :math:`\vect{\theta} \in \Rset^{n_\theta}`.
Under the Gaussian process prior assumption, the observations
:math:`\vect{Y} = \left(Y_i, i = 1, \ldots, m\right)` and a prediction
:math:`Y(\vect{x})` at some unobserved input :math:`\vect{x}` are
jointly normally distributed:
.. math::
\begin{pmatrix}
\vect{Y} \\
Y(\vect{x})
\end{pmatrix}
\thicksim
\cN_{m + 1}\left(
\begin{pmatrix}
\mat{F} \vect{\beta} \\
\Tr{\vect{f}(\vect{x})} \vect{\beta}
\end{pmatrix}
,\,\sigma^2
\begin{pmatrix}
\mat{R} & \vect{r}(\vect{x}) \\
\vect{r}(\vect{x})^t & 1
\end{pmatrix}
\right)
where:
.. math:: \mat{F} = \left[f_j\left(\vect{x}^{(i)}\right), i = 1, \ldots, m, j = 1, \ldots, p\right]
is the regression matrix,
.. math:: \mat{R} = \left[R\left(\vect{x}^{(i)} - \vect{x}^{(j)}, \vect{\theta}\right), i,\,j = 1, \ldots, m\right]
is the observations’ correlation matrix, and:
.. math:: \vect{r}(\vect{x}) = \Tr{\left(R\left(\vect{x} - \vect{x}^{(i)}, \vect{\theta}\right), i = 1, \ldots, m\right)}
is the vector of cross-correlations between the prediction and the
observations.
As such, the Kriging predictor is defined as the following conditional
distribution:
.. math:: \tilde{Y}(\vect{x}) = \left[Y(\vect{x}) \mid \vect{Y} = \vect{y}, \vect{\theta}=\vect{\theta}^*, \sigma^2=(\sigma^2)^*\right]
where :math:`\vect{\theta}^*` and :math:`(\sigma^2)^*` are the maximum
likelihood estimates of the correlation parameters :math:`\vect{\theta}`
and variance :math:`\sigma^2` (see references).
It can be shown (see references) that the predictor
:math:`\tilde{Y}(\vect{x})` is also Gaussian:
.. math:: \tilde{Y}(\vect{x}) = \cN_1\left(\mu_{\tilde{Y}}(\vect{x}), \sigma^2_{\tilde{Y}}(\vect{x})\right)
- with mean:
.. math:: \mu_{\tilde{Y}}(\vect{x}) = \Tr{\vect{f}(\vect{x})} \tilde{\vect{\beta}} + \Tr{\vect{r}(\vect{x})} \vect{\gamma}
where :math:`\underline{\tilde{\beta}}` is the generalized least
squares solution of the underlying linear regression problem:
.. math:: \tilde{\vect{\beta}} = \left(\Tr{\mat{F}} \mat{R}^{-1} \mat{F}\right)^{-1} \Tr{\mat{F}} \mat{R}^{-1} \vect{y}
and
.. math:: \vect{\gamma} = \mat{R}^{-1} \left(\vect{y} - \mat{F} \tilde{\vect{\beta}}\right)
- and variance:
.. math::
\sigma^2_{\tilde{Y}}(\vect{x}) =
\sigma^2 \left[1 -
\Tr{\vect{r}(\vect{x})} \mat{R}^{-1} \vect{r}(\vect{x})
+ \Tr{\vect{u}(\vect{x})} \left(\Tr{\mat{F}} \mat{R}^{-1} \mat{F}\right)^{-1} \vect{u}(\vect{x})
\right]
where:
.. math:: \vect{u}(\vect{x}) = \Tr{\mat{F}} \mat{R}^{-1} \vect{r}(\vect{x}) - \vect{f}(\vect{x})
Kriging may also be referred to as *Gaussian process regression*.
.. topic:: API:
- See :class:`~openturns.KrigingAlgorithm`
.. topic:: Examples:
- See :doc:`/auto_meta_modeling/kriging_metamodel/plot_kriging`
.. topic:: References:
- [dubourg2011]_
- S. Lophaven, H. Nielsen and J. Sondergaard, 2002, "DACE, A Matlab kriging toolbox", Technichal University of Denmark. https://www.omicron.dk/dace/dace.pdf
- T. Santner, B. Williams and W. Notz, 2003. "The design and analysis of computer experiments", Springer, New York.
- C. Rasmussen and C. Williams, 2006, T. Dietterich (Ed.), "Gaussian processes for machine learning", MIT Press.
|
