基于监督神经网络的宽吸收线类星体星表
天体物理学
2009-11-13 v1
摘要
我们应用学习向量量化(LVQ)算法处理SDSS DR5类星体光谱,以构建一个大型的宽吸收线类星体(BALQSO)星表。首先讨论了使用传统balnicity指数和/或吸收指数(BI和AI)构建BALQSO星表时存在的问题,然后描述了为识别BALQSO而训练的监督LVQ网络。由此得到的BALQSO星表在鲁棒性和完备性上应显著优于基于BI或AI的星表。
引用
@article{arxiv.0810.4396,
title = {Broad Absorption Line Quasar catalogues with Supervised Neural Networks},
author = {Simone Scaringi and Christopher E. Cottis and Christian Knigge and Michael R. Goad},
journal= {arXiv preprint arXiv:0810.4396},
year = {2009}
}
备注
5 pages, 3 figures, to appear in the proceedings of "Classification and Discovery in Large Astronomical Surveys", Ringberg Castle, 14-17 October 2008