Abstract
In this chapter we introduce an improved parallel optimal choropleth map classification algorithm to support spatial analysis. This work contributes to the development of a Distributed Geospatial CyberInfrastructure and offers an implementation of the Fisher-Jenks optimal classification method suitable for multi-core desktop environments. We provide a description of both a single-core vectorized implementation and a parallelized implementation. Our results show that single core vectorization alone provides computational speedups compared to previous parallel implementations and that a combined, parallel and vectorized, implementation offers significant speed improvements.
Original language | English (US) |
---|---|
Title of host publication | Modern accelerator technologies for geographic information science |
Publisher | Springer US |
Pages | 197-212 |
Number of pages | 16 |
Volume | 9781461487456 |
ISBN (Electronic) | 9781461487456 |
ISBN (Print) | 1461487447, 9781461487449 |
DOIs | |
State | Published - Aug 1 2013 |
Keywords
- Parallelization •
- PySAL
- Spatial analysis •
- Vectorization •
ASJC Scopus subject areas
- General Computer Science