|
N. N

06.11.2008, 17:00 Uhr s.t. im M5 Vortrag von Professor Dr. Horst Bunke, Universität Bern unter dem Titel Recent Developments in Graph Classification and Clustering Using Graph Embedding Kernels

Veröffentlicht Friday, 17.10.2008 11:49

Mathematik und Informatik

Abstract: Graphs provide us with a powerful and flexible representation formalism for pattern recognition. However, the vast majority of pattern recognition algorithms rely on vectorial data descriptions and cannot directly be applied to graphs. Recently a growing interest in graph kernel methods can be observed. Graph kernels aim at bridging the gap between the high representational power and flexibility of graphs and the large amount of algorithms available for object representations based on feature vectors. In this talk we review recent work that aims at transforming graphs into n-dimensional real vectors by means of prototype selection and graph edit distance computation. This approach allows one to build graph kernels in a straightforward way. With several classification and clustering experiments we prove the robustness and flexibility of our new method and show that this approach outperforms other graph classification and clustering methods on several graph data sets of diverse nature.



Angelegt am Friday, 17.10.2008 11:49 von N. N
Geändert am Friday, 17.10.2008 11:49 von N. N
[Edit | Vorlage]

Kolloquium der Informatik