XGBoost classifier filters interlopers in CSST slitless spectroscopy simulations, retaining 42% of galaxies with 96.6% accurate redshifts and 0.13% outliers.
Unbiased Cosmological Parameter Estimation from Emission Line Surveys with Interlopers
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
abstract
The galaxy catalogs generated from low-resolution emission line surveys often contain both foreground and background interlopers due to line misidentification, which can bias the cosmological parameter estimation. In this paper, we present a method for correcting the interloper bias by using the joint-analysis of auto- and cross-power spectra of the main and the interloper samples. In particular, we can measure the interloper fractions from the cross-correlation between the interlopers and survey galaxies, because the true cross-correlation must be negligibly small. The estimated interloper fractions, in turn, remove the interloper bias in the cosmological parameter estimation. For example, in the Hobby-Eberly Telescope Dark Energy Experiment (HETDEX) low-redshift ($z<0.5$) [O II] $\lambda3727${\AA} emitters contaminate high-redshift ($1.9<z<3.5$) Lyman-$\alpha$ line emitters. We demonstrate that the joint-analysis method yields a high signal-to-noise ratio measurement of the interloper fractions while only marginally increasing the uncertainties in the cosmological parameters relative to the case without interlopers. We also show the same is true for the high-latitude spectroscopic survey of Wide-Field Infrared Survey Telescope (WFIRST) mission where contamination occurs between the Balmer-$\alpha$ line emitters at lower redshifts ($1.1<z<1.9$) and Oxygen ([O III] $\lambda5007${\AA}) line emitters at higher redshifts ($1.7<z<2.8$).
representative citing papers
HETDEX PDR1 releases an expanded public catalog of 426,654 LAEs, 491,411 [O II] emitters and additional sources from 86.67 sq deg of integral-field spectroscopy with improved classifiers.
Constructs a (D_τ,D_x)-manifold with N-correlators of N_t-objects using field theory, topology, algebra, statistics and Fourier transforms, and discusses applicability across cosmological scales.
citing papers explorer
-
Filtering Interlopers with Photometry and Diagnostic Features: A Machine Learning Framework Validated with CSST Slitless Spectroscopy
XGBoost classifier filters interlopers in CSST slitless spectroscopy simulations, retaining 42% of galaxies with 96.6% accurate redshifts and 0.13% outliers.
-
HETDEX Public Data Release 1: Source Catalog 2 and Data Cubes from ~90 sq deg of Integral-Field Optical Spectroscopy
HETDEX PDR1 releases an expanded public catalog of 426,654 LAEs, 491,411 [O II] emitters and additional sources from 86.67 sq deg of integral-field spectroscopy with improved classifiers.
-
A $(D_\tau,D_x)$-manifold with $N$-correlators of $N_t$-objects
Constructs a (D_τ,D_x)-manifold with N-correlators of N_t-objects using field theory, topology, algebra, statistics and Fourier transforms, and discusses applicability across cosmological scales.