Abstract
Given a spatial point pattern, we use various characteristics of its Voronoi diagram and Delaunay tessellation to extract information of the dependence between points. In particular, we use the characteristics to construct statistics for testing complete spatial randomness. It is shown that the minimum angle of a typical Delaunay triangle is sensitive to both regularity and clustering alternatives, whilst the triangle's area or perimeter is more sensitive to clustering than regularity. These statistics are also sensitive to the Baddeley-Silverman cell process.
Original language | English |
---|---|
Pages (from-to) | 367-376 |
Number of pages | 10 |
Journal | Biometrical Journal |
Volume | 45 |
Issue number | 3 |
DOIs | |
Publication status | Published - Apr 2003 |
Scopus Subject Areas
- Statistics and Probability
- Statistics, Probability and Uncertainty
User-Defined Keywords
- Complete spatial randomness
- Delaunay tessellation
- Goodness of fit
- Spatial point pattern
- Voronoi diagram