Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
473712 | Computers & Mathematics with Applications | 2011 | 14 Pages |
Abstract
We propose a new approach, based on the Conley index theory, for the detection and classification of critical regions in multidimensional data sets. The use of homology groups makes this method consistent and successful in all dimensions and allows us to generalize visual classification techniques based solely on the notion of connectedness which may fail in higher dimensions.
Related Topics
Physical Sciences and Engineering
Computer Science
Computer Science (General)
Authors
Madjid Allili, David Corriveau, Sara Derivière, Marc Ethier, Tomasz Kaczynski,