Cluster Pattern Detection in Spatial Data Based on Monte Carlo Inference

Clusters in a data point field exhibit spatially specified regions in the observation window. The method proposed in this paper addresses the cluster detection problem from the perspective of detection of these spatial regions. These regions are supposed to be formed of overlapping random disks driven by a marked point process. The distribution of such a process has two components. The first is related to the location of the disks in the field of observation and is defined as an inhomogeneous Poisson process. The second one is related to the interaction between disks and is constructed by the superposition of an area-interaction and a pairwise interaction processes. The model is applied on spatial data coming from animal epidemiology. The proposed method tackles several aspects related to cluster pattern detection: heterogeneity of data, smoothing effects, statistical descriptors, probability of cluster presence, testing for the cluster presence.

[1]  P. Green,et al.  Boundary detection through dynamic polygons , 1998 .

[2]  M. N. M. van Lieshout,et al.  Perfect simulation for marked point processes , 2006, Comput. Stat. Data Anal..

[3]  van Marie-Colette Lieshout,et al.  Stochastic annealing for nearest-neighbour point processes with application to object recognition , 1994, Advances in Applied Probability.

[4]  Josiane Zerubia,et al.  A Gibbs Point Process for Road Extraction from Remotely Sensed Images , 2004, International Journal of Computer Vision.

[5]  W. Kendall,et al.  Perfect simulation using dominating processes on ordered spaces, with application to locally stable point processes , 2000, Advances in Applied Probability.

[6]  B. Ripley,et al.  Markov Point Processes , 1977 .

[7]  Josiane Zerubia,et al.  Point processes for unsupervised line network extraction in remote sensing , 2005, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[8]  Jesper Møller,et al.  Bayesian contour detection in a time series of ultrasound images through dynamic deformable template models. , 2002, Biostatistics.

[9]  J. Møller,et al.  An efficient Markov chain Monte Carlo method for distributions with intractable normalising constants , 2006 .

[10]  Håvard Rue,et al.  Bayesian object recognition with baddeley's delta loss , 1998, Advances in Applied Probability.

[11]  Josiane Zerubia,et al.  A RJMCMC Algorithm for Object Processes in Image Processing , 2001, Monte Carlo Methods Appl..

[12]  Merrilee Hurn,et al.  Bayesian object identification , 1999 .

[13]  A. Baddeley,et al.  Area-interaction point processes , 1993 .

[14]  Håvard Rue,et al.  Identification of partly destroyed objects using deformable templates , 1998, Stat. Comput..

[15]  Jorge Mateu,et al.  Simulated annealing and object point processes: Tools for analysis of spatial patterns , 2005 .

[16]  C. Geyer,et al.  Simulation Procedures and Likelihood Inference for Spatial Point Processes , 1994 .

[17]  P. Green.  Reversible jump Markov chain Monte Carlo computation and Bayesian model determination , 1995 .