| |
 |
| Paper: |
Neural Networks for Estimating Galaxy Redshifts from a Multi-Band Photometric Astronomical Survey |
| Volume: |
522, Astronomical Data Analysis Software and Systems XXVII |
| Page: |
405 |
| Authors: |
Muralikrishna, A.; Junior, W. A. d. S.; dos Santos, R. D. C. |
| Abstract: |
The quest for accuracy in galaxy maps of the Universe is ever more important for studies on cosmological models. To that end, Machine Learning methods have been most helpful, with rather good accuracy, on estimating variables, such as photometric redshifts. Combined with Data Science concepts, works on astronomical Big Data with good results can be processed by a research group in a clear and reproducible way, which can easily aid further work development and/or optimize results already obtained by other research groups. This present work focus on using a promising tool to estimate photometric redshifts: Artificial Neural Networks. Although they have been used already, we plan to eventually develop this approach further, by utilizing state-of-the-art Deep Learning methods. The preliminary results for photometric redshifts from neural networks we obtained seem promising, with a σNMAD value of 2.2%. Besides that, applying Data Science concepts we also aim to create and make public available a documentation with all steps required to reproduce and enhance the results we obtained. |
|
|
|
|
|
