Professional Documents
Culture Documents
SH Schmidt
SH Schmidt
SollerhausWorkshop,
March 2006
Introduction
Aims
1. Explain some basic ideas for stochastic modelling of
spatial point patterns
stationarity (homogeneity) vs. spatial trends
isotropy (rotational invariance)
complete spatial randomness
interaction between points (clustering, repulsion)
SollerhausWorkshop,
March 2006
Introduction
Aims
1. Explain some basic ideas for stochastic modelling of
spatial point patterns
stationarity (homogeneity) vs. spatial trends
isotropy (rotational invariance)
complete spatial randomness
interaction between points (clustering, repulsion)
2. Describe some basic characteristics of pointprocess
models
Intensity measure, conditional intensities
Pair correlation function, Ripleys K-function, etc.
SollerhausWorkshop,
March 2006
Introduction
Aims
1. Explain some basic ideas for stochastic modelling of
spatial point patterns
stationarity (homogeneity) vs. spatial trends
isotropy (rotational invariance)
complete spatial randomness
interaction between points (clustering, repulsion)
2. Describe some basic characteristics of pointprocess
models
Intensity measure, conditional intensities
Pair correlation function, Ripleys K-function, etc.
3. Present techniques for the statistical analysis of spatial
point patterns
SollerhausWorkshop,
March 2006
Introduction
Overview
1. Examples of spatial point patterns
SollerhausWorkshop,
March 2006
Introduction
Overview
1. Examples of spatial point patterns
point patterns in networks
SollerhausWorkshop,
March 2006
Introduction
Overview
1. Examples of spatial point patterns
point patterns in networks
other point patterns
SollerhausWorkshop,
March 2006
Introduction
Overview
1. Examples of spatial point patterns
point patterns in networks
other point patterns
2. Models for spatial point patterns
SollerhausWorkshop,
March 2006
Introduction
Overview
1. Examples of spatial point patterns
point patterns in networks
other point patterns
2. Models for spatial point patterns
Poisson point processes
SollerhausWorkshop,
March 2006
Introduction
Overview
1. Examples of spatial point patterns
point patterns in networks
other point patterns
2. Models for spatial point patterns
Poisson point processes
Cluster and hardcore processes
SollerhausWorkshop,
March 2006
Introduction
Overview
1. Examples of spatial point patterns
point patterns in networks
other point patterns
2. Models for spatial point patterns
Poisson point processes
Cluster and hardcore processes
Gibbs point processes
SollerhausWorkshop,
March 2006
Introduction
Overview
1. Examples of spatial point patterns
point patterns in networks
other point patterns
2. Models for spatial point patterns
Poisson point processes
Cluster and hardcore processes
Gibbs point processes
3. Characteristics of point processes
SollerhausWorkshop,
March 2006
Introduction
Overview
1. Examples of spatial point patterns
point patterns in networks
other point patterns
2. Models for spatial point patterns
Poisson point processes
Cluster and hardcore processes
Gibbs point processes
3. Characteristics of point processes
4. Some statistical issues
SollerhausWorkshop,
March 2006
Introduction
Overview
1. Examples of spatial point patterns
point patterns in networks
other point patterns
2. Models for spatial point patterns
Poisson point processes
Cluster and hardcore processes
Gibbs point processes
3. Characteristics of point processes
4. Some statistical issues
Nonparametric estimation of model characteristics
SollerhausWorkshop,
March 2006
Introduction
Overview
1. Examples of spatial point patterns
point patterns in networks
other point patterns
2. Models for spatial point patterns
Poisson point processes
Cluster and hardcore processes
Gibbs point processes
3. Characteristics of point processes
4. Some statistical issues
Nonparametric estimation of model characteristics
Maximum pseudolikelihood estimation
SollerhausWorkshop,
March 2006
Examples
Point patterns in networks
SollerhausWorkshop,
March 2006
Examples
Random tessellations
Poisson-Voronoi (PVT)
Poisson-Delaunay (PDT)
PLT/PLT
SollerhausWorkshop,
March 2006
PLT/PDT
PVT/PLT
5
Examples
Model fitting for network data
Cutout
Tessellation model
(PLT/PLT)
SollerhausWorkshop,
March 2006
Examples
Analysis of biological networks
b) Lamellipodium
SollerhausWorkshop,
March 2006
Examples
Analysis of biological networks
Graph structure
Tessellation model
(PVT/PLT)
SollerhausWorkshop,
March 2006
Graph structure
Tessellation model
(PLT/PDT)
8
Other examples
Modelling of tropical storm tracks
1945-2004
SollerhausWorkshop,
March 2006
Other examples
Point patterns in biological cell nuclei
Heterochromatin structures in interphase nuclei
SollerhausWorkshop,
March 2006
10
Other examples
Point patterns in biological cell nuclei
SollerhausWorkshop,
March 2006
11
Other examples
Point patterns in biological cell nuclei
Capsides of
cytomegalovirus
SollerhausWorkshop,
March 2006
12
Models
Basic ideas
1. Mathematical definition of spatial point processes
Let {X1 , X2 , ...} be a sequence of random vectors
with values in 2 and
let X(B) = #{n : Xn B} denote the number of
points Xn located in a set B 2 .
SollerhausWorkshop,
March 2006
13
Models
Basic ideas
1. Mathematical definition of spatial point processes
Let {X1 , X2 , ...} be a sequence of random vectors
with values in 2 and
let X(B) = #{n : Xn B} denote the number of
points Xn located in a set B 2 .
If X(B) < for each bounded set B 2 , then
{X1 , X2 , ...} is called a random point process.
SollerhausWorkshop,
March 2006
13
Models
Basic ideas
1. Mathematical definition of spatial point processes
Let {X1 , X2 , ...} be a sequence of random vectors
with values in 2 and
let X(B) = #{n : Xn B} denote the number of
points Xn located in a set B 2 .
If X(B) < for each bounded set B 2 , then
{X1 , X2 , ...} is called a random point process.
SollerhausWorkshop,
March 2006
13
Models
Basic ideas
1. Mathematical definition of spatial point processes
Let {X1 , X2 , ...} be a sequence of random vectors
with values in 2 and
let X(B) = #{n : Xn B} denote the number of
points Xn located in a set B 2 .
If X(B) < for each bounded set B 2 , then
{X1 , X2 , ...} is called a random point process.
SollerhausWorkshop,
March 2006
13
Models
Basic ideas
1. Stochastic model vs. single realization
SollerhausWorkshop,
March 2006
14
Models
Basic ideas
1. Stochastic model vs. single realization
Point processes are mathematical models
SollerhausWorkshop,
March 2006
14
Models
Basic ideas
1. Stochastic model vs. single realization
Point processes are mathematical models
Observed point patterns are their realizations
SollerhausWorkshop,
March 2006
14
Models
Basic ideas
1. Stochastic model vs. single realization
Point processes are mathematical models
Observed point patterns are their realizations
2. Some further remarks
Equivalent notions:
point field instead of spatial point process
SollerhausWorkshop,
March 2006
14
Models
Basic ideas
1. Stochastic model vs. single realization
Point processes are mathematical models
Observed point patterns are their realizations
2. Some further remarks
Equivalent notions:
point field instead of spatial point process
Point processes are not necessarily dynamic
SollerhausWorkshop,
March 2006
14
Models
Basic ideas
1. Stochastic model vs. single realization
Point processes are mathematical models
Observed point patterns are their realizations
2. Some further remarks
Equivalent notions:
point field instead of spatial point process
Point processes are not necessarily dynamic
Dynamics (w.r.t. time/space) can be added
=> spatial birth-and-death processes
SollerhausWorkshop,
March 2006
14
Models
Stationary Poisson process
1. Definition of stationary Poisson processes
Poisson distribution of point counts:
X(B) Poi(|B|) for any bounded B
>0
SollerhausWorkshop,
March 2006
and some
15
Models
Stationary Poisson process
1. Definition of stationary Poisson processes
Poisson distribution of point counts:
X(B) Poi(|B|) for any bounded B
>0
and some
SollerhausWorkshop,
March 2006
15
Models
Stationary Poisson process
1. Definition of stationary Poisson processes
Poisson distribution of point counts:
X(B) Poi(|B|) for any bounded B
>0
and some
2. Basic properties
X(B) = |B| = = intensity of points
Voidprobabilities: (X(B) = 0) = exp(|B|)
SollerhausWorkshop,
March 2006
15
Models
Stationary Poisson process
1. Definition of stationary Poisson processes
Poisson distribution of point counts:
X(B) Poi(|B|) for any bounded B
>0
and some
2. Basic properties
X(B) = |B| = = intensity of points
Voidprobabilities: (X(B) = 0) = exp(|B|)
SollerhausWorkshop,
March 2006
15
Models
Stationary Poisson process
SollerhausWorkshop,
March 2006
16
Models
Stationary Poisson process
SollerhausWorkshop,
March 2006
16
Models
Stationary Poisson process
SollerhausWorkshop,
March 2006
16
Models
Stationary Poisson process
SollerhausWorkshop,
March 2006
16
Models
General Poisson process
1. Definition of general (nonhomogeneous) Poisson
processes
Poisson distribution of point counts:
X(B) Poi((B)) for any bounded B 2 and
some measure : B( 2 ) [0, ]
SollerhausWorkshop,
March 2006
17
Models
General Poisson process
1. Definition of general (nonhomogeneous) Poisson
processes
Poisson distribution of point counts:
X(B) Poi((B)) for any bounded B 2 and
some measure : B( 2 ) [0, ]
SollerhausWorkshop,
March 2006
17
Models
General Poisson process
1. Definition of general (nonhomogeneous) Poisson
processes
Poisson distribution of point counts:
X(B) Poi((B)) for any bounded B 2 and
some measure : B( 2 ) [0, ]
2. Basic properties
X(B) = (B) = = intensity measure
SollerhausWorkshop,
March 2006
17
Models
General Poisson process
1. Definition of general (nonhomogeneous) Poisson
processes
Poisson distribution of point counts:
X(B) Poi((B)) for any bounded B 2 and
some measure : B( 2 ) [0, ]
2. Basic properties
X(B) = (B) = = intensity measure
SollerhausWorkshop,
March 2006
17
Models
General Poisson process
1. Definition of general (nonhomogeneous) Poisson
processes
Poisson distribution of point counts:
X(B) Poi((B)) for any bounded B 2 and
some measure : B( 2 ) [0, ]
2. Basic properties
X(B) = (B) = = intensity measure
SollerhausWorkshop,
March 2006
17
SollerhausWorkshop,
March 2006
18
SollerhausWorkshop,
March 2006
18
SollerhausWorkshop,
March 2006
18
SollerhausWorkshop,
March 2006
19
SollerhausWorkshop,
March 2006
19
SollerhausWorkshop,
March 2006
19
SollerhausWorkshop,
March 2006
19
SollerhausWorkshop,
March 2006
19
SollerhausWorkshop,
March 2006
20
SollerhausWorkshop,
March 2006
20
SollerhausWorkshop,
March 2006
20
SollerhausWorkshop,
March 2006
21
SollerhausWorkshop,
March 2006
21
SollerhausWorkshop,
March 2006
21
SollerhausWorkshop,
March 2006
21
SollerhausWorkshop,
March 2006
21
SollerhausWorkshop,
March 2006
22
SollerhausWorkshop,
March 2006
22
SollerhausWorkshop,
March 2006
22
2. Special cases
stationary Poisson process: (u, x) =
SollerhausWorkshop,
March 2006
22
2. Special cases
stationary Poisson process: (u, x) =
SollerhausWorkshop,
March 2006
22
2. Special cases
stationary Poisson process: (u, x) =
SollerhausWorkshop,
March 2006
22
SollerhausWorkshop,
March 2006
23
SollerhausWorkshop,
March 2006
23
SollerhausWorkshop,
March 2006
23
SollerhausWorkshop,
March 2006
23
SollerhausWorkshop,
March 2006
23
SollerhausWorkshop,
March 2006
23
SollerhausWorkshop,
March 2006
24
SollerhausWorkshop,
March 2006
24
SollerhausWorkshop,
March 2006
24
SollerhausWorkshop,
March 2006
24
SollerhausWorkshop,
March 2006
24
SollerhausWorkshop,
March 2006
24
SollerhausWorkshop,
March 2006
25
b)
a)
Initial configuration
SollerhausWorkshop,
March 2006
Birth of a point
25
b)
a)
Initial configuration
Birth of a point
d)
c)
Birth of another point
SollerhausWorkshop,
March 2006
Death of a point
25
SollerhausWorkshop,
March 2006
26
SollerhausWorkshop,
March 2006
26
2. Properties:
SollerhausWorkshop,
March 2006
26
2. Properties:
Minimal interpoint distance r (hardcore radius r/2)
SollerhausWorkshop,
March 2006
26
2. Properties:
Minimal interpoint distance r (hardcore radius r/2)
(r, R) = interval of interaction distances
(attraction/clustering if > 1, repulsion if < 1)
SollerhausWorkshop,
March 2006
26
2. Properties:
Minimal interpoint distance r (hardcore radius r/2)
(r, R) = interval of interaction distances
(attraction/clustering if > 1, repulsion if < 1)
SollerhausWorkshop,
March 2006
26
SollerhausWorkshop,
March 2006
27
SollerhausWorkshop,
March 2006
27
Initial configuration
SollerhausWorkshop,
March 2006
Death of a point
SollerhausWorkshop,
March 2006
28
SollerhausWorkshop,
March 2006
28
SollerhausWorkshop,
March 2006
28
= 0.01
SollerhausWorkshop,
March 2006
= 0.1
28
SollerhausWorkshop,
March 2006
29
SollerhausWorkshop,
March 2006
29
SollerhausWorkshop,
March 2006
29
and
SollerhausWorkshop,
March 2006
VarX(B) =
(2) (B
B) + (B) (B)
2
29
and
VarX(B) =
(2) (B
R
B) + (B) (B)
R
2
(2)
Often
1 B2 ) = B1 B2 (u1 , u2 ) du1 du2
(2) (u1 , u2 ) = product density
(2) (B
SollerhausWorkshop,
March 2006
29
and
VarX(B) =
(2) (B
R
B) + (B) (B)
R
2
(2)
Often
1 B2 ) = B1 B2 (u1 , u2 ) du1 du2
(2) (u1 , u2 ) = product density
(2) (B
SollerhausWorkshop,
March 2006
29
and
VarX(B) =
(2) (B
R
B) + (B) (B)
R
2
(2)
Often
1 B2 ) = B1 B2 (u1 , u2 ) du1 du2
(2) (u1 , u2 ) = product density
(2) (B
SollerhausWorkshop,
March 2006
29
SollerhausWorkshop,
March 2006
30
SollerhausWorkshop,
March 2006
30
SollerhausWorkshop,
March 2006
30
SollerhausWorkshop,
March 2006
30
SollerhausWorkshop,
March 2006
30
SollerhausWorkshop,
March 2006
30
SollerhausWorkshop,
March 2006
30
where
SollerhausWorkshop,
March 2006
2 K(B)
B1
K(B2 u) du
(2) (x) dx
31
where
2 K(B)
B1
K(B2 u) du
(2) (x) dx
SollerhausWorkshop,
March 2006
31
where
2 K(B)
B1
K(B2 u) du
(2) (x) dx
SollerhausWorkshop,
March 2006
31
where
2 K(B)
B1
K(B2 u) du
(2) (x) dx
SollerhausWorkshop,
March 2006
(= |Bs (o)|)
31
where
2 K(B)
B1
K(B2 u) du
(2) (x) dx
(= |Bs (o)|)
SollerhausWorkshop,
March 2006
31
SollerhausWorkshop,
March 2006
s>0
32
s>0
D(s) = 1 lim
0
SollerhausWorkshop,
March 2006
32
s>0
D(s) = 1 lim
0
SollerhausWorkshop,
March 2006
32
s>0
D(s) = 1 lim
0
SollerhausWorkshop,
March 2006
32
s>0
D(s) = 1 lim
0
SollerhausWorkshop,
March 2006
32
SollerhausWorkshop,
March 2006
33
SollerhausWorkshop,
March 2006
33
SollerhausWorkshop,
March 2006
33
Properties:
b is unbiased, i.e.,
b=
SollerhausWorkshop,
March 2006
33
Properties:
b is unbiased, i.e.,
b=
b with probability 1
If {Xn } is ergodic, then
b is strongly consistent
(as |W | ), i.e.
SollerhausWorkshop,
March 2006
33
SollerhausWorkshop,
March 2006
34
SollerhausWorkshop,
March 2006
34
SollerhausWorkshop,
March 2006
2 K(s)
(2) (x) dx
Bs (o)
35
2 K(s)
(2) (x) dx
Bs (o)
SollerhausWorkshop,
March 2006
35
2 K(s)
(2) (x) dx
Bs (o)
SollerhausWorkshop,
March 2006
35
2 K(s)
(2) (x) dx
Bs (o)
SollerhausWorkshop,
March 2006
35
Xi ,Xj W, i6=j
where k :
SollerhausWorkshop,
March 2006
k(|Xi Xj | s)
|(W + Xi ) (W + Xj )|
36
Xi ,Xj W, i6=j
where k :
k(|Xi Xj | s)
|(W + Xi ) (W + Xj )|
SollerhausWorkshop,
March 2006
36
Xi ,Xj W, i6=j
where k :
k(|Xi Xj | s)
|(W + Xi ) (W + Xj )|
SollerhausWorkshop,
March 2006
36
SollerhausWorkshop,
March 2006
37
b :
Nearest-neighbor distance distribution function D(s)
b
D(s)
=
SollerhausWorkshop,
March 2006
Xn W
37
b :
Nearest-neighbor distance distribution function D(s)
b
D(s)
=
Xn W
SollerhausWorkshop,
March 2006
37
SollerhausWorkshop,
March 2006
38
SollerhausWorkshop,
March 2006
38
n
Y
i=1
SollerhausWorkshop,
March 2006
Z
(xi , x) exp
(u, x) du
38
n
Y
i=1
Z
(xi , x) exp
(u, x) du
SollerhausWorkshop,
March 2006
38
SollerhausWorkshop,
March 2006
j=1
39
j=1
SollerhausWorkshop,
March 2006
39
j=1
SollerhausWorkshop,
March 2006
39
j=1
SollerhausWorkshop,
March 2006
39
j=1
SollerhausWorkshop,
March 2006
39
m
X
j=1
m
X
j=1
wj (yj log j j )
SollerhausWorkshop,
March 2006
40
m
X
j=1
m
X
j=1
wj (yj log j j )
SollerhausWorkshop,
March 2006
40
m
X
j=1
m
X
j=1
wj (yj log j j )
SollerhausWorkshop,
March 2006
40
References
A. Baddeley, P. Gregori, J. Mateu, R. Stoica, D. Stoyan (eds.): Case Studies in
Spatial Point Process Modeling. Lecture Notes in Statistics 185, Springer, New York
2006.
J. Moeller, R.P. Waagepetersen: Statistical Inference and Simulation for Spatial Point
Processes. Monographs on Statistics and Applied Probability 100, Chapman and
Hall, Boca Raton 2004.
D. Stoyan, W.S. Kendall, J. Mecke: Stochastic Geometry and Its Applications
(2nd ed.). J. Wiley & Sons, Chichester 1995.
D. Stoyan, H. Stoyan: Fractals, Random Shapes and Point Fields. J. Wiley & Sons,
Chichester 1994.
SollerhausWorkshop,
March 2006
41