Galaxy classification: deep learning on the OTELO and COSMOS databases

Abstract

Context. The accurate classification of hundreds of thousands of galaxies observed in modern deep surveys is imperative if we want to understand the universe and its evolution. Aims. Here, we report the use of machine learning techniques to classify early- and late-type galaxies in the OTELO and COSMOS databases using optical and infrared photometry and available shape parameters: either the Sérsic index or the concentration index. Methods. We used three classification methods for the OTELO database: (1) u − r color separation, (2) linear discriminant analysis using u − r and a shape parameter classification, and (3) a deep neural network using the r magnitude, several colors, and a shape parameter. We analyzed the performance of each method by sample bootstrapping and tested the performance of our neural network architecture using COSMOS data. Results. The accuracy achieved by the deep neural network is greater than that of the other classification methods, and it can also operate with missing data. Our neural network architecture is able to classify both OTELO and COSMOS datasets regardless of small differences in the photometric bands used in each catalog. Conclusions. In this study we show that the use of deep neural networks is a robust method to mine the cataloged data.

The OTELO survey as a morphological probe. Last ten Gyr of galaxy evolution

“…Dot-dashed lines show the results of linear discriminant analysis of both colours with (V GF − I GF ) = 1.7 and (u−r) = 2.5. same panel of Fig. 10, we show the results of the linear discriminant analysis from de Diego et al (2020). Using this discriminant for the morphological sample, we found 98% and 66% of completeness for LT and ET, respectively.…”

“…We are aware of the possible misclassification of LT and ET using SED templates. However, de Diego et al (2020) showed that less than 2% are misclassified using dense neural networks for the subset of the data used in this work (see their Sect. 3.1.4).…”

“…The morphological analysis is important for a full exploration of the possible scientific cases of the OTELO survey. Among them, we can highlight the following: a census of ET galaxies with emission lines, a comprehensive study of the properties of compact galaxies, a comparison of extragalactic sources with and without detection of emission lines; and a recently published work on the machine learning techniques to separate ET from LT galaxies (de Diego et al 2020). Additionally, in this work, we study the MSR of galaxies up to z = 2 and down to stellar masses of log M * /M ∼ 8, which is ∼1 dex lower than the lower mass limit established in previous studies at the same redshift range (van der Wel et al 2014;Mowla et al 2019).…”

The OTELO survey

“…The red dashed line marks the linear discriminant analysis developed by de Diego et al (2020), employing deep learning methods for morphology classification. In the figure we can see that all the selected emitters (indicated by red circles) are comfortably distributed in the late-type locus.…”

Nonsequential neural network for simultaneous, consistent classification, and photometric redshifts of OTELO galaxies

“…Machine learning aims to make accurate predictions after learning how to map input data to outputs. The most popular machine learning techniques employed in astronomy either for morphological classification, which we refer to as just classification, or redshift estimates are Support Vector Machines (SVM; e.g., Wadadekar 2005;Jones & Singal 2017;Khramtsov et al 2020) and particularly the GalSVM code (Huertas-Company et al 2008;Pović et al 2012Pović et al , 2013Pović et al , 2015Pintos-Castro et al 2016;Amado et al 2019), Random Forests (e.g., Carrasco Kind & Brunner 2013Mucesh et al 2021), and neural networks (e.g., Serra-Ricart et al 1993Firth et al 2003;Domínguez Sánchez et al 2018;de Diego et al 2020). All these are supervised techniques; they need a labeled dataset for training.…”