Robust SOM-SuperSOM Clustering

Authors:
Joan Perez*
mail: joan.perez@univ-cotedazur.fr

Giovanni Fusco*
mail: giovanni.fusco@univ-cotedazur.fr

*Univ. Côte d'Azur, CNRS, ESPACE, UMR7300
98 Bd Herriot, BP 3209, 06204 Nice, France

This release allows performing a combined SOM/SuperSOM clustering of the 640 administrative districts of India.
Fowlkes-Mallows Similarity Index is used to identify robust initializations of the clustering.
Data_India.txt is a specially conceived geographic database of 55 indicators, covering issues of economic activity,
urban structure, socio-demographic development, consumption levels, infrastructure endowment and
basic geographical positioning within the Indian space. Data refer to 2011 or to 2001-2011 evolutions.
A shapefile of the Indian district is attached allowing to map the results.

v.1.1 contains

• R Script (Robust SOM Clustering v1.1)
• Data set (Data_India.txt)
• Archive of shapefiles for the Indian districts of 2011 (District_2011.zip)
• Fusco G., Perez J., 2015, Spatial Analysis of the Indian Subcontinent: the Complexity Investigated through Neural Networks
• README.txt

References:

[1] Wehrens R., Buydens L., 2007, Self- and Super-organizing Maps in R: The ko-honen Package,
Journal of Statistical Software, Vol. 21, Issue 5.
[2] Fusco G., Perez J., 2015, Spatial Analysis of the Indian Subcontinent: the Complexity Investigated
through Neural Networks, CUPUM 2015 - 14th International Conference on Computers in Urban Planning
and Urban Management, MIT, Cambridge (Ma.), July 5th-7th 2015, Proceedings, 287, 1-20,
http://web.mit.edu/cron/project/CUPUM2015/proceedings/Content/analytics/287_fusco_h.pdf
[3] Perez J., 2015, Spatial Structures in India in the Age of Globalisation. A Data-Driven Approach, Phd in geography,
University of Avignon (France)