| |
 |
| Paper: |
Applications of Support Vector Machines in Astronomy |
| Volume: |
485, Astronomical Data Analysis Software and Systems XXIII |
| Page: |
239 |
| Authors: |
Zhang, Y.; Zhao, Y. |
| Abstract: |
We review Support Vector Machines (SVMs) as applied in astronomy. SVMs
are mainly used for solving the and regression
issues. Take classification for example, selecting of cataclysmic
variables from large spectroscopic survey, detecting quasar
candidates from multiwavelength photometric data, identification of
blue horizontal branch stars from photometric data, classification
of galactic spectra, supernova search; for regression problem,
photometric redshift estimation of galaxies and quasars, physical
parameter measurement (metallicity, gravity, effective temperature)
of stars. Comparatively, SVMs show better performance in
classification than in regression. Nevertheless, SVMs has its
disadvantages, which needs large computation cost on training. Based
on this problem, CUDA-Accelerated SVMs is put forward. As for
accuracy of SVMs, SVMs combined with other algorithms has further
improvement, such as SVM-KNN. |
|
|
|
|
|
