| Paper: | MLSP-P6.11 | ||
| Session: | Biomedical and Other Applications | ||
| Time: | Friday, May 19, 16:30 - 18:30 | ||
| Presentation: | Poster | ||
| Topic: | Machine Learning for Signal Processing: Other Applications | ||
| Title: | On Dimensionality Reduction for Classification and its Application | ||
| Authors: | Raviv Raich, University of Michigan, United States; Jose Costa, California Institute of Technology, United States; Alfred O. Hero, III, University of Michigan, United States | ||
| Abstract: | In this paper, we evaluate the contribution of the classification constrained dimensionality reduction (CCDR) algorithm to the performance of several classifiers. We present an extension to previously introduced CCDR algorithm to multiple hypotheses. We investigate classification performance using the CCDR algorithm on hyper-spectral satellite imagery data. We demonstrate the performance gain for both local and global classifiers and demonstrate a $10\%$ improvement of the $k$-nearest neighbors algorithm performance. We present a connection between intrinsic dimension estimation and the optimal embedding dimension obtained using the CCDR algorithm. | ||
©2018 Conference Management Services, Inc. -||- email: webmaster@icassp2006.org -||- Last updated Friday, August 17, 2012