Euclidean Distances, Soft and Spectral Clustering on Weighted Graphs
نویسنده
چکیده
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.
منابع مشابه
On Symmetry of Some Nano Structures
It is necessary to generate the automorphism group of a chemical graph in computer-aided structure elucidation. An Euclidean graph associated with a molecule is defined by a weighted graph with adjacency matrix M = [dij], where for i≠j, dij is the Euclidean distance between the nuclei i and j. In this matrix dii can be taken as zero if all the nuclei are equivalent. Otherwise, one may introduce...
متن کاملHeat Kernels, Manifolds and Graph Embedding
In this paper, we investigate the use of heat kernels as a means of embedding graphs in a pattern space. We commence by performing the spectral decomposition on the graph Laplacian. The heat kernel of the graph is found by exponentiating the resulting eigensystem over time. By equating the spectral heat kernel and its Gaussian form we are able to approximate the geodesic distance between nodes ...
متن کاملAUTOMORPHISM GROUPS OF SOME NON-TRANSITIVE GRAPHS
An Euclidean graph associated with a molecule is defined by a weighted graph with adjacency matrix M = [dij], where for ij, dij is the Euclidean distance between the nuclei i and j. In this matrix dii can be taken as zero if all the nuclei are equivalent. Otherwise, one may introduce different weights for distinct nuclei. Balaban introduced some monster graphs and then Randic computed complexit...
متن کاملSpectral Analysis of Signed Graphs for Clustering, Prediction and Visualization
We study the application of spectral clustering, prediction and visualization methods to graphs with negatively weighted edges. We show that several characteristic matrices of graphs can be extended to graphs with positively and negatively weighted edges, giving signed spectral clustering methods, signed graph kernels and network visualization methods that apply to signed graphs. In particular,...
متن کاملMinimum Weight Euclidean Matching and Weighted Relative Neighborhood Graphs
The Minimum Weight Euclidean Matching (MWEM) problem is: given 2n point sites in the plane with Euclidean metric for interpoint distances, match the sites into n pairs so that the sum of the n distances between matched pairs is minimized. The graph theoretic version of this problem has been extensively studied since the pioneering work of Edmonds. The best time bound known for MEWM is O(n 2:5 (...
متن کامل