Unbiased clustering estimation in the presence of missing observations
Unbiased clustering estimation in the presence of missing observations
复制标题
DOI：
10.1093/mnras/stx2053
复制
发表时间：
2017-03
影响因子：
4.8
通讯作者：
D. Bianchi;W. Percival
中科院分区：
文献类型：
--
作者：
D. Bianchi;W. Percival
为了提高效率，光谱星系红移调查并不能获得目标群体中所有星系的红移。失踪的星系通常是聚集在一起的，这通常导致在密集区域成功观测的比例较低。一个例子是SDSS光谱星系巡天的近对问题，它有一个观测星系对的缺陷，它们的角间距比放置相邻纤维的硬件限制更近。空间聚集的缺失观测将在下一代的调查中存在。以前已经提出了各种方案来减轻这些影响，但没有一种方案适用于所有情况。我们认为，解决方案是将失踪的星系与观测到的具有统计等效聚类属性的星系联系起来，而做到这一点的最佳方法是重新运行目标算法，改变观测的角度位置。假设每个配对在一个算法实现中被观察到的概率非零，那么将目标与成功观察联系起来的配对提升方案可以纠正这些问题。我们提出了这样一个方案，并使用一个理想化的简单调查策略的实现来证明其有效性。
In order to be efficient, spectroscopic galaxy redshift surveys do not obtain redshifts for all galaxies in the population targeted. The missing galaxies are often clustered, commonly leading to a lower proportion of successful observations in dense regions. One example is the close-pair issue for SDSS spectroscopic galaxy surveys, which have a deficit of pairs of observed galaxies with angular separation closer than the hardware limit on placing neighbouring fibers. Spatially clustered missing observations will exist in the next generations of surveys. Various schemes have previously been suggested to mitigate these effects, but none works for all situations. We argue that the solution is to link the missing galaxies to those observed with statistically equivalent clustering properties, and that the best way to do this is to rerun the targeting algorithm, varying the angular position of the observations. Provided that every pair has a non-zero probability of being observed in one realisation of the algorithm, then a pair-upweighting scheme linking targets to successful observations, can correct these issues. We present such a scheme, and demonstrate its validity using realisations of an idealised simple survey strategy.