-
Augmenting machine learning photometric redshifts with Gaussian mixture models
Authors:
P. W. Hatfield,
I. A. Almosallam,
M. J. Jarvis,
N. Adams,
R. A. A. Bowler,
Z. Gomes,
S. J. Roberts,
C. Schreiber
Abstract:
Wide-area imaging surveys are one of the key ways of advancing our understanding of cosmology, galaxy formation physics, and the large-scale structure of the Universe in the coming years. These surveys typically require calculating redshifts for huge numbers (hundreds of millions to billions) of galaxies - almost all of which must be derived from photometry rather than spectroscopy. In this paper…
▽ More
Wide-area imaging surveys are one of the key ways of advancing our understanding of cosmology, galaxy formation physics, and the large-scale structure of the Universe in the coming years. These surveys typically require calculating redshifts for huge numbers (hundreds of millions to billions) of galaxies - almost all of which must be derived from photometry rather than spectroscopy. In this paper we investigate how using statistical models to understand the populations that make up the colour-magnitude distribution of galaxies can be combined with machine learning photometric redshift codes to improve redshift estimates. In particular we combine the use of Gaussian Mixture Models with the high performing machine learning photo-z algorithm GPz and show that modelling and accounting for the different colour-magnitude distributions of training and test data separately can give improved redshift estimates, reduce the bias on estimates by up to a half, and speed up the run-time of the algorithm. These methods are illustrated using data from deep optical and near infrared data in two separate deep fields, where training and test data of different colour-magnitude distributions are constructed from the galaxies with known spectroscopic redshifts, derived from several heterogeneous surveys.
△ Less
Submitted 3 September, 2020;
originally announced September 2020.
-
Non-Gaussianity Constraints using Future Radio Continuum Surveys and the Multi-Tracer Technique
Authors:
Zahra Gomes,
Stefano Camera,
Matt J. Jarvis,
Catherine Hale,
José Fonseca
Abstract:
Tighter constraints on measurements of primordial non-Gaussianity will allow the differentiation of inflationary scenarios. The cosmic microwave background bispectrum-the standard method of measuring the local non-Gaussianity-is limited by cosmic variance. Therefore, it is sensible to investigate measurements of non-Gaussianity using the large-scale structure. This can be done by investigating the…
▽ More
Tighter constraints on measurements of primordial non-Gaussianity will allow the differentiation of inflationary scenarios. The cosmic microwave background bispectrum-the standard method of measuring the local non-Gaussianity-is limited by cosmic variance. Therefore, it is sensible to investigate measurements of non-Gaussianity using the large-scale structure. This can be done by investigating the effects of non-Gaussianity on the power spectrum on large scales. In this study we forecast the constraints on the local primordial non-Gaussianity parameter $f_{\rm NL}$ that can be obtained with future radio surveys. We utilize the multi-tracer method which reduces the effect of cosmic variance and takes advantage of the multiple radio galaxy populations which are differently biased tracers of the same underlying dark matter distribution. Improvements on previous work include the use of observational bias and halo mass estimates, updated simulations and realistic photometric redshift expectations, thus producing more realistic forecasts. Combinations of SKA simulations and radio observations were used as well as different redshift ranges and redshift bin sizes. It was found that in the most realistic case the 1 - $σ$ error on $f_{\rm NL}$ falls within the range 4.07 and 6.58, rivalling the tightest constraints currently available.
△ Less
Submitted 17 December, 2019;
originally announced December 2019.
-
Tomographic galaxy clustering with the Subaru Hyper Suprime-Cam first year public data release
Authors:
Andrina Nicola,
David Alonso,
Javier Sánchez,
Anže Slosar,
Humna Awan,
Adam Broussard,
Jo Dunkley,
Zahra Gomes,
Eric Gawiser,
Rachel Mandelbaum,
Hironao Miyatake,
Jeffrey A. Newman,
Ignacio Sevilla,
Sarah Skinner,
Erica Wagoner
Abstract:
We analyze the clustering of galaxies in the first public data release of the HSC Subaru Strategic Program. Despite the relatively small footprints of the observed fields, the data are an excellent proxy for the deep photometric datasets that will be acquired by LSST, and are therefore an ideal test bed for the analysis methods being implemented by the LSST DESC. We select a magnitude limited samp…
▽ More
We analyze the clustering of galaxies in the first public data release of the HSC Subaru Strategic Program. Despite the relatively small footprints of the observed fields, the data are an excellent proxy for the deep photometric datasets that will be acquired by LSST, and are therefore an ideal test bed for the analysis methods being implemented by the LSST DESC. We select a magnitude limited sample with $i<24.5$ and analyze it in four redshift bins covering $0.15\lesssim z \lesssim1.5$. We carry out a Fourier-space analysis of the two-point clustering of this sample, including all auto- and cross-correlations. We demonstrate the use of map-level deprojection methods to account for fluctuations in the galaxy number density caused by observational systematics. Through an HOD analysis, we place constraints on the characteristic halo masses of this sample, finding a good fit up to scales $k_{\rm max}=1\,{\rm Mpc}^{-1}$, including both auto- and cross-correlations. Our results show monotonically decreasing average halo masses, which can be interpreted in terms of the drop-out of red galaxies at high redshifts for a flux-limited sample. In terms of photometric redshift systematics, we show that additional care is needed in order to marginalize over uncertainties in the redshift distribution in galaxy clustering, and that these uncertainties can be constrained by including cross-correlations. We are able to make a $\sim3σ$ detection of lensing magnification in the HSC data. Our results are stable to variations in $σ_8$ and $Ω_c$ and we find constraints that agree well with measurements from Planck and low-redshift probes. Finally, we use our pipeline to study the clustering of galaxies as a function of limiting flux, and provide a simple fitting function for the linear galaxy bias for magnitude limited samples as a function of limiting magnitude and redshift. [abridged]
△ Less
Submitted 17 December, 2019;
originally announced December 2019.
-
Improving Photometric Redshift Estimation using GPz: size information, post processing and improved photometry
Authors:
Zahra Gomes,
Matt J. Jarvis,
Ibrahim A. Almosallam,
Stephen J. Roberts
Abstract:
The next generation of large scale imaging surveys (such as those conducted with the Large Synoptic Survey Telescope and Euclid) will require accurate photometric redshifts in order to optimally extract cosmological information. Gaussian Processes for photometric redshift estimation (GPz) is a promising new method that has been proven to provide efficient, accurate photometric redshift estimations…
▽ More
The next generation of large scale imaging surveys (such as those conducted with the Large Synoptic Survey Telescope and Euclid) will require accurate photometric redshifts in order to optimally extract cosmological information. Gaussian Processes for photometric redshift estimation (GPz) is a promising new method that has been proven to provide efficient, accurate photometric redshift estimations with reliable variance predictions. In this paper, we investigate a number of methods for improving the photometric redshift estimations obtained using GPz (but which are also applicable to others). We use spectroscopy from the Galaxy and Mass Assembly Data Release 2 with a limiting magnitude of r<19.4 along with corresponding Sloan Digital Sky Survey visible (ugriz) photometry and the UKIRT Infrared Deep Sky Survey Large Area Survey near-IR (YJHK) photometry. We evaluate the effects of adding near-IR magnitudes and angular size as features for the training, validation and testing of GPz and find that these improve the accuracy of the results by ~15-20 per cent. In addition, we explore a post-processing method of shifting the probability distributions of the estimated redshifts based on their Quantile-Quantile plots and find that it improves the bias by ~40 per cent. Finally, we investigate the effects of using more precise photometry obtained from the Hyper Suprime-Cam Subaru Strategic Program Data Release 1 and find that it produces significant improvements in accuracy, similar to the effect of including additional features.
△ Less
Submitted 6 December, 2017;
originally announced December 2017.
