Machine learning in APOGEE

Garcia‐Dias, Rafael; Prieto, C. Allende; Alméida, J. Sánchez; Ordovás-Pascual, I.

doi:10.1051/0004-6361/201732134

Cited by 22 publications

(6 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The project is collecting hundreds of thousands of stellar spectra with high signal-to-noise ratios across the Milky Way, focusing on the regions where dust causes dramatic extinction at optical wavelengths, namely the disk and the central parts of the Galaxy. The APOGEE observations are providing a chemical map of our Galaxy with unprecedented quality (Badenes et al 2018;Fernández-Alvar et al 2018;Fu et al 2018;García-Dias et al 2018;Garcııa-Pérez et al 2018;Hayes et al 2018;Mackereth et al 2019;Palicio et al 2018;Souto et al 2018;Weinberg et al 2018;Wilson et al 2018).…”

Section: Introductionmentioning

confidence: 99%

Radial Velocities in the Outermost Disk toward the Anticenter

López-Corredoira

Labini

Kalberla

et al. 2019

View full text Add to dashboard Cite

We measure the mean Galactocentric radial component of the velocity of stars (v R ) in the disk at 8 kpc< R < 28 kpc in the direction of the anticenter. For this, we use the Apache Point Galactic Evolution Experiment (APOGEE). Furthermore, we compare the result with HI maps along the same line of sight. We find an increase in positive (expansion) v R at R ≈ 9 − 13 kpc, reaching a maximum of ≈ 6 km/s, and a decrease at large values of R reaching a negative (contraction) value of ≈ −10 km/s for R > 17 kpc. Negative velocities are also observed in 21 cm HI maps, possibly dominated by local gas emission. Among the possible dynamical causes for these non-zero v R , factors such as the effect of the Galactic bar, streams, or mergers do not seem appropriate to explain our observations. An explanation might be the gravitational attraction of overdensities in a spiral arm. As a matter of fact, we see a change of regime from positive to negative velocities around R ≈ 15 kpc, in the position where we cross the Outer spiral arm in the anticenter. The mass in spiral arms necessary to produce these velocities would be about 3% of the mass of the disk, consistent with our knowledge of the spiral arms. Another scenario that we explore is a simple class of out-of-equilibrium systems in which radial motions are generally created by the monolithic collapse of isolated self-gravitating overdensities.

show abstract

Section: Introductionmentioning

confidence: 99%

Radial Velocities in the Outermost Disk toward the Anticenter

López-Corredoira

Labini

Kalberla

et al. 2019

View full text Add to dashboard Cite

show abstract

“…The v-measure score and homogeneity score are transparent to permutations of the labels, but the accuracy score needs all the clusters to be cross-matched. In this case, we matched each group of stars found by the unsupervised tool to the star cluster with the highest number of member stars inside the group, as was done in Sánchez Almeida & Allende Prieto (2013) and Garcia- Dias et al (2018). However, in this work when the number of clusters in the real dataset did not match the number of clusters in the predicted model, or when the objects in one group did not match any of the available clusters, we assigned the group to the cluster with the highest number of coincident objects, even when the cluster had previously been assigned to another group.…”

Section: Clustering Algorithmsmentioning

confidence: 99%

“…For example, supervised spectral classification has been adopted in works such as those by Bailer-Jones et al (1998); Singh et al (1998); Bailer-Jones (2001); Rodríguez et al (2004); Giridhar et al (2006); Manteiga et al (2009), andNavarro et al (2012). Unsupervised spectral classification was also explored in works such as Sánchez Almeida Bovy (2017); Traven et al (2017); Valentini et al (2017); Garcia- Dias et al (2018); Reis et al (2018), and Price-Jones & Bovy (2019).…”

Section: Introductionmentioning

confidence: 99%

Machine learning in APOGEE: Identification of stellar populations through chemical abundances

Garcia-Dias,

Prieto,

Almeida

et al. 2019

Preprint

Self Cite

View full text Add to dashboard Cite

Context. The vast volume of data generated by modern astronomical surveys offers test beds for the application of machine-learning. In these exploratory applications, it is important to evaluate potential existing tools and determine those that are optimal for extracting scientific knowledge from the available observations. Aims. We explore the possibility of using unsupervised clustering algorithms to separate stellar populations with distinct chemical patterns. Methods. Star clusters are likely the most chemically homogeneous populations in the Galaxy, and therefore any practical approach to identifying distinct stellar populations should at least be able to separate clusters from each other. We have applied eight clustering algorithms combined with four dimensionality reduction strategies to automatically distinguish stellar clusters using chemical abundances of 13 elements. Our test-bed sample includes 18 stellar clusters with a total of 453 stars. Results. We have applied statistical tests showing that some pairs of clusters (e.g., NGC 2458-NGC 2420) are indistinguishable from each other when chemical abundances from the Apache Point Galactic Evolution Experiment (APOGEE) are used. However, for most clusters we are able to automatically assign membership with metric scores similar to previous works. The confusion level of the automatically selected clusters is consistent with statistical tests that demonstrate the impossibility of perfectly distinguishing all the clusters from each other. These statistical tests and confusion levels establish a limit for the prospect of blindly identifying stars born in the same cluster based solely on chemical abundances. Conclusions. We find that some of the algorithms we explored are capable of blindly identify stellar populations with similar ages and chemical distributions in the APOGEE data. Even though we are not able to fully separate the clusters from each other, the main confusion arises from clusters with similar ages. Because some stellar clusters are chemically indistinguishable, our study supports the notion of extending weak chemical tagging that involves families of clusters instead of individual clusters.

show abstract

“…Among all classification methods, K-means clustering has been the most commonly used in analyzing astronomical data, e.g., searching for extremely metal-poor galaxies (Sánchez Almeida & Allende Prieto 2013), classification of stellar spectra obtained by SDSS/SEGUE and APOGEE (Sánchez Almeida et al 2016;Garcia-Dias et al 2018). Classical methods including K-means belong to "hard classification," which adopts a nonprobabilistic model and deduces the classification result from the decision function, where the sample is definitely assigned to one class.…”

Section: Introductionmentioning

confidence: 99%

Fuzzy Cluster Analysis: Application to Determining Metallicities for Very Metal-poor Stars

2021

ApJ

View full text Add to dashboard Cite

This work presents a first attempt to apply fuzzy cluster analysis (FCA) to analyzing stellar spectra. FCA is adopted to categorize line indices measured from LAMOST low-resolution spectra, and automatically remove the least metallicity-sensitive indices. The FCA-processed indices are then transferred to the artificial neural network (ANN) to derive metallicities for 147 very metal-poor (VMP) stars that have been analyzed by high-resolution spectroscopy. The FCA-ANN method could derive robust metallicities for VMP stars, with a precision of ∼0.2 dex compared with high-resolution analysis. The recommended FCA threshold value λ for this test is between 0.9965 and 0.9975. After reducing the dimension of the line indices through FCA, the derived metallicities are still robust, with no loss of accuracy, and the FCA-ANN method performs stably for different spectral quality from [Fe/H] ∼ −1.8 down to −3.5. Compared with traditional classification methods, FCA considers ambiguity in groupings and noncontinuity of data, and is thus more suitable for observational data analysis. Though this early test uses FCA to analyze low-resolution spectra, and feeds the input to the ANN method to derive metallicities, FCA should be able to, in the large data era, also analyze slitless spectroscopy and multiband photometry, and prepare the input for methods not limited to ANN, in the field of stellar physics for other studies, e.g., stellar classification, identification of peculiar objects. The literature-collected high-resolution sample can help improve pipelines to derive stellar metallicities, and systematic offsets in metallicities for VMP stars for three published LAMOST catalogs have been discussed.

show abstract

Machine learning in APOGEE

Cited by 22 publications

References 40 publications

Radial Velocities in the Outermost Disk toward the Anticenter

Radial Velocities in the Outermost Disk toward the Anticenter

Machine learning in APOGEE: Identification of stellar populations through chemical abundances

Fuzzy Cluster Analysis: Application to Determining Metallicities for Very Metal-poor Stars

Contact Info

Product

Resources

About