Euclidean Distances, soft and spectral Clustering on Weighted Graphs
Machine Learning
2010-09-15 v1
Abstract
We define a class of Euclidean distances on weighted graphs, enabling to perform thermodynamic soft graph clustering. The class can be constructed form the "raw coordinates" encountered in spectral clustering, and can be extended by means of higher-dimensional embeddings (Schoenberg transformations). Geographical flow data, properly conditioned, illustrate the procedure as well as visualization aspects.
Cite
@article{arxiv.1007.0832,
title = {Euclidean Distances, soft and spectral Clustering on Weighted Graphs},
author = {François Bavaud},
journal= {arXiv preprint arXiv:1007.0832},
year = {2010}
}
Comments
accepted for presentation (and further publication) at the ECML PKDD 2010 conference