Application of Kohonen Self-organizing Map for urban structure analysis
Document Type
Conference Proceeding
Publication Date
11-22-2006
Abstract
Kohonen Self-organizing Map (SOM) has been widely used to discover clusters in datasets of various real world applications. Urban structure, as characterized by various social, economic, and environmental features, can be explored with SOM. In this paper, we present a case study of applying SOM for urban structure analysis to the city of New Orleans. One novel aspect of this work is the inclusion of environmental data from satellite imagery for clustering. Ten social-economic variables from Census 2000 data and Normalized Difference Vegetation Index (NDVI) from Landsat 7 Enhanced Thematic Mapper Plus (ETM+) imagery were used as inputs to SOM to group 482 census block groups of New Orleans into 9 clusters. The clustering results show that block groups with high economic status and environmental quality tended to cluster at 3 major outer locations, while block groups with low economic status and environmental quality were concentrated in the mid-city area, which is a well known pattern due to the suburbanization process. Three major components were extracted by a principal component analysis, and they were compared with the clustering results from SOM. Fisher's linear discriminant analysis shows a good separability result among 9 clusters discovered by the SOM. © 2006 IEEE.
Publication Source (Journal or Book title)
2006 IEEE International Conference on Granular Computing
First Page
118
Last Page
123
Recommended Citation
Ju, W., Lam, N., & Chen, J. (2006). Application of Kohonen Self-organizing Map for urban structure analysis. 2006 IEEE International Conference on Granular Computing, 118-123. Retrieved from https://repository.lsu.edu/eecs_pubs/2410