Unsupervised classification in high dimension
Abstract
Dealing with large databases of galaxy spectra is a good example of a new problematic task in astrophysics. Current and forthcoming big surveys provide millions of spectra each containing thousands of wavelengths. These spectra must be confronted with physical and chemical models. This requires an unsupervised classification which is a dimensionality reduction in both the number of observations and parameters. In this poster, we present some approaches that we are implementing.
| Origin | Files produced by the author(s) |
|---|
Loading...
