Multi-Classification by Categorical Features via Clustering
We derive a generalization bound for multi-classification schemes based on grid clustering in categorical parameter product spaces. Grid clustering partitions the parameter space in the form of a Cartesian product of partitions for each of the parameters. The derived bound provides a means to evaluate clustering solutions in terms of the generalization power of a built-on classifier. For classification based on a single feature the bound serves to find a globally optimal classification rule. Comparison of the generalization power of individual features can then be used for feature ranking. Our experiments show that in this role the bound is much more precise than mutual information or normalized correlation indices.
| Author(s): | Seldin, Y. and Tishby, N. |
| Journal: | In the proceedings of the 25th International Conference on Machine Learning (ICML 2008) |
| Pages: | 920-927 |
| Year: | 2008 |
| Month: | June |
| Day: | 0 |
| Bibtex Type: | Conference Paper (inproceedings) |
| Event Name: | 25th International Conference on Machine Learning (ICML 2008) |
| Digital: | 0 |
| Electronic Archiving: | grant_archive |
| Language: | en |
| Organization: | Max-Planck-Gesellschaft |
| School: | Biologische Kybernetik |
| Links: | |
BibTex
@inproceedings{6575,
title = {Multi-Classification by Categorical Features via Clustering},
journal = {In the proceedings of the 25th International Conference on Machine Learning (ICML 2008)},
abstract = {We derive a generalization bound for multi-classification schemes based on grid clustering in categorical parameter product spaces. Grid clustering partitions the parameter space in the form of a Cartesian product of partitions for each of the parameters. The derived bound provides a means to evaluate clustering solutions in terms of the generalization power of a built-on classifier. For classification based on a single feature the bound serves to find a globally optimal classification rule. Comparison of the generalization power of individual features can then be used for feature ranking. Our experiments show that in this role the bound is much more precise than mutual information or normalized correlation indices.},
pages = {920-927},
organization = {Max-Planck-Gesellschaft},
school = {Biologische Kybernetik},
month = jun,
year = {2008},
slug = {6575},
author = {Seldin, Y. and Tishby, N.},
month_numeric = {6}
}