CORSO is a graph-partitioning algorithm to capture spatial continuity of some envi- ronment, that is, it detects spatially close regions which are sharing similar properties. Areal units (or graph nodes) are described in terms of properties (e.g., extension, temperature) of the areal units as well as properties or relations (e.g. distance, parallelism) involving the primary units (e.g. rivers, streets) collected within each area boundary. Closeness is modeled by means of spatial con- straints (graph edges), such as adjacency. CORSO resorts to an ILP approach to mine relational data and exploits the concept of neighborhood of a seed object to capture relational constraints. Cluster shape depends on similarity evaluation and seed selection criteria. Solutions to face both these issues are illustrated in this work.
Homogeneity Evaluation and Seed Selection in Clustering Graph-Connected Spatial Data
APPICE, ANNALISA;LANZA, Antonietta;MALERBA, Donato;
2007-01-01
Abstract
CORSO is a graph-partitioning algorithm to capture spatial continuity of some envi- ronment, that is, it detects spatially close regions which are sharing similar properties. Areal units (or graph nodes) are described in terms of properties (e.g., extension, temperature) of the areal units as well as properties or relations (e.g. distance, parallelism) involving the primary units (e.g. rivers, streets) collected within each area boundary. Closeness is modeled by means of spatial con- straints (graph edges), such as adjacency. CORSO resorts to an ILP approach to mine relational data and exploits the concept of neighborhood of a seed object to capture relational constraints. Cluster shape depends on similarity evaluation and seed selection criteria. Solutions to face both these issues are illustrated in this work.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
