en
Proceedings chapter
Open access
English

Unsupervised Discriminant Embedding in Cluster Spaces

Presented at Madeira (Portugal), Oct 6-8, 2009
Publication date2009
Abstract

This paper proposes a new representation space, called the cluster space, for data points that originate from high dimensions. Whereas existing dedicated methods concentrate on revealing manifolds from within the data, we consider here the context of clustered data and derive the dimension reduction process from cluster information. Points are represented in the cluster space by means of their a posteriori probability values estimated using Gaussian Mixture Models. The cluster space obtained is the optimal space for discrimination in terms of the Quadratic Discriminant Analysis (QDA).Moreover, it is shown to alleviate the negative impact of the curse of dimensionality on the quality of cluster discrimination and is a useful preprocessing tool for other dimension reduction methods. Various experiments illustrate the effectiveness of the cluster space bothon synthetic and real data.

Keywords
  • Dimension reduction
  • Clustering
  • High dimensionality
Citation (ISO format)
SZEKELY, Eniko-Melinda, BRUNO, Eric, MARCHAND-MAILLET, Stéphane. Unsupervised Discriminant Embedding in Cluster Spaces. In: International Conference on Knowledge Discovery and Information Retrieval (KDIR′09). Madeira (Portugal). [s.l.] : [s.n.], 2009.
Main files (1)
Proceedings chapter (Accepted version)
accessLevelPublic
Identifiers
  • PID : unige:47654
433views
270downloads

Technical informations

Creation03/06/2015 5:12:05 PM
First validation03/06/2015 5:12:05 PM
Update time03/14/2023 10:58:24 PM
Status update03/14/2023 10:58:24 PM
Last indexation01/16/2024 5:07:38 PM
All rights reserved by Archive ouverte UNIGE and the University of GenevaunigeBlack