arXiv subjects

M. Catelan

Publications and source records attributed to M. Catelan.

At least 37 records · Page 2Linked to original sources

arXiv2024

A study of nine extreme low mass ratio contact binary systems

Low mass ratio systems (LMR) are a very interesting class of contact eclipsing binaries challenging the theoretical models of stability. These systems are also considered possible progenitors of the rare low-mass optical transients called red novae. In this study, we present the identification of 7 new totally eclipsing LMR systems from Catalina Sky Surveys (CSS) and 77 LMR candidates from the All Sky Automated Survey (ASAS$-$3). Using the available CSS light curves and new multiband observations for CSS$\_$J210228.3$-$031048 and CSS$\_$J231513.$+$345335 with the 2.3 m Aristarchos telescope at Helmos Observatory, we estimate their physical and absolute parameters and investigate their stability and their progenitors. The light curves are analyzed by performing a 2$-$dimension scan on the mass ratio $-$ inclination plane with Phoebe$-$0.31 scripter while the errors are estimated using Monte-Carlo simulations and heuristic scanning of the parameter space. Our analysis revealed that all 9 CSS systems have extreme mass ratios from 0.09 to 0.16. Our statistical analysis of well$-$studied LMR contact binaries shows that LMR systems tend to have warmer and more massive primaries. The investigation of the progenitors of both low and higher-mass ratio systems reveals a trend for the former to originate from higher-mass ancestors. Finally, we investigate the stability condition by calculating the ratio of spin angular momentum to orbital angular momentum and other stability indicators in the context of the reliability of the solutions.

astro-ph.SR↗

arXiv2024

ATAT: Astronomical Transformer for time series And Tabular data

The advent of next-generation survey instruments, such as the Vera C. Rubin Observatory and its Legacy Survey of Space and Time (LSST), is opening a window for new research in time-domain astronomy. The Extended LSST Astronomical Time-Series Classification Challenge (ELAsTiCC) was created to test the capacity of brokers to deal with a simulated LSST stream. We describe ATAT, the Astronomical Transformer for time series And Tabular data, a classification model conceived by the ALeRCE alert broker to classify light-curves from next-generation alert streams. ATAT was tested in production during the first round of the ELAsTiCC campaigns. ATAT consists of two Transformer models that encode light curves and features using novel time modulation and quantile feature tokenizer mechanisms, respectively. ATAT was trained on different combinations of light curves, metadata, and features calculated over the light curves. We compare ATAT against the current ALeRCE classifier, a Balanced Hierarchical Random Forest (BHRF) trained on human-engineered features derived from light curves and metadata. When trained on light curves and metadata, ATAT achieves a macro F1-score of 82.9 +- 0.4 in 20 classes, outperforming the BHRF model trained on 429 features, which achieves a macro F1-score of 79.4 +- 0.1. The use of Transformer multimodal architectures, combining light curves and tabular data, opens new possibilities for classifying alerts from a new generation of large etendue telescopes, such as the Vera C. Rubin Observatory, in real-world brokering scenarios.

astro-ph.IM↗

arXiv2024

Spectroscopic confirmation of high-amplitude eruptive YSOs and dipping giants from the VVV survey

During the pre-main-sequence (pre-MS) evolution stage of a star, significant amounts of stellar mass are accreted during episodic accretion events, such as multi-decade FUor-type outbursts. Here, we present a near-infrared spectroscopic follow-up study of 33 high-amplitude (most with $ΔK_s$ > 4 mag) variable sources discovered by the Vista Variables in the Via Lactea (VVV) survey. Based on the spectral features, 25 sources are classified as eruptive young stellar objects (YSOs), including 15 newly identified FUors, six with long-lasting but EXor-like bursts of magnetospheric accretion and four displaying outflow-dominated spectra. By examining the photometric behaviours of eruptive YSOs, we found most FUor-type outbursts have higher amplitudes ($ΔK_s$ and $ΔW2$), faster eruptive timescales and bluer infrared colours than the other outburst types. In addition, we identified seven post-main sequence variables apparently associated with deep dipping events and an eruptive star with deep AlO absorption bands resembling those seen in the V838 Mon stellar merger.

astro-ph.SR↗

arXiv2024

The most variable VVV sources: eruptive protostars, dipping giants in the Nuclear Disc and others

We have performed a comprehensive search of a VISTA Variables in the Via Lactea (VVV) database of 9.5 yr light curves for variable sources with $ΔK_s \ge 4$ mag, aiming to provide a large sample of high amplitude eruptive young stellar objects (YSOs) and detect unusual or new types of infrared variable source. We find 222 variable or transient sources in the Galactic bulge and disc, most of which are new discoveries. The sample mainly comprises novae, YSOs, microlensing events, Long Period Variable stars (LPVs) and a few rare or unclassified sources. Additionally, we report the discovery of a significant population of aperiodic late-type giant stars suffering deep extinction events, strongly clustered in the Nuclear Disc of the Milky Way. We suggest that these are metal-rich stars in which radiatively driven mass loss has been enhanced by super-solar metallicity. Among the YSOs, 32/40 appear to be undergoing episodic accretion. Long-lasting YSO eruptions have a typical rise time of $\sim$2 yr, somewhat slower than the 6-12 month timescale seen in the few historical events observed on the rise. The outburst durations are usually at least 5 yr, somewhat longer than many lower amplitude VVV events detected previously. The light curves are diverse in nature, suggesting that multiple types of disc instability may occur. Eight long-duration extinction events are seen wherein the YSO dims for a year or more, attributable to inner disc structure. One binary YSO in NGC 6530 displays periodic extinction events (P=59 days) similar to KH 15D.

astro-ph.SR↗

arXiv2023

Period Change Rates of Large Magellanic Cloud Cepheids using MESA

Pulsating stars, such as Cepheids and RR Lyrae, offer us a window to measure and study changes due to stellar evolution. In this work, we study the former by calculating a set of evolutionary tracks of stars with an initial mass of 4 to 7 $M_\odot$, varying the initial rotation rate and metallicity, using the stellar evolution code Modules for Experiments in Stellar Astrophysics (MESA). Using Radial Stellar Pulsations (RSP), a recently added functionality of MESA, we obtained theoretical instability strip (IS) edges and linear periods for the radial fundamental mode. Period-age, period-age-temperature, period-luminosity, and period-luminosity-temperature relationships were derived for three rotation rates and metallicities, showing a dependence on crossing number, position in the IS, rotation, and metallicity. We calculated period change rates (PCRs) based on the linear periods from RSP. We compared our models with literature results using the Geneva code, and found large differences, as expected due to the different implementations of rotation between codes. In addition, we compared our theoretical PCRs with those measured in our recent work for Large Magellanic Cloud Cepheids. We found good overall agreement, even though our models do not reach the short-period regime exhibited by the empirical data. Implementations of physical processes not yet included in our models, such as pulsation-driven mass loss, an improved treatment of convection that may lead to a better description of the instability strip edges, as well as consideration of a wider initial mass range, could all help improve the agreement with the observed PCRs.

astro-ph.SR↗

arXiv2023

A Unique Low-Mass Ratio Contact Eclipsing Binary System under the Period Cut-Off

We present a multi-band photometric analysis of CRTS J163819.6+03485, the first low mass ratio (LMR) contact binary system with a period under the contact binary (CB) period limit. The unprecedented combination of mass ratio and period makes this system unique for eclipsing binary (EB) research. Using new multi-band photometric observations, we explored the parameter space of this unique total EB system through a detailed scan in the mass ratio - inclination plane and using the PIKAIA genetic algorithm optimizer. The best set of relative physical parameters and corresponding uncertainties was adopted through Markov Chain Monte Carlo sampling of the parameter space. The resulting mass ratio of the system is $q = 0.16 \pm 0.01$. The absolute parameters were derived by adopting an empirical mass-luminosity relation. Period changes are also investigated by using new observations and archival photometric light curves from massive astronomical surveys, which revealed in a preliminary solution the presence of a possible low-mass tertiary companion. The origin and evolutionary status of the system are investigated through the detached-binary formation scenario.

astro-ph.SR↗

arXiv2023

Uncertainties in the 12C+12C reaction rate and their impact on the composition of ultra-massive WDs

Stars with initial masses 7 Msun . MZAMS . 9 Msun reach temperatures high enough to ignite C under degenerate conditions after the end of He-core burning (Garcia-Berro & Iben 1994). These isolated stars are expected to evolve into the so-called super AGB (SAGB) phase and may end their lives as ultra-massive ONe WDs (see Siess 2006,2007, 2010; Camisassa et al. 2019, and references therein). The exact proportions of O and Ne found in the core at the end of the SAGB phase will determine the cooling times and pulsational properties of these WDs. Uncertainties affecting the rates of nuclear reactions occurring during the C burning phase should have a measurable impact on the distribution of 16O, 20Ne, 23Na and 24Mg and, consequently, on the evolution of the WD. Here we present a study of the impact of uncertainties in the 12C(12C, α)20Ne and 12C(12C, p)23Na nuclear reaction rates (and their branching ratios) on the chemical structure of intermediate- to high-mass progenitors at the end of the C-burning phase. Using the stellar evolution code Modules for Experiments in Stellar Astrophysics (MESA) we computed evolutionary sequences for stars with initial masses 7.25<= MZAMS /Msun <=8.25, from the ZAMS to the SAGB phase, adopting different prescriptions for the 12C+12C burning rates. We found that adopting lower reaction rates for the 12C+12C burning delays C-ignition by at most 2700 yrs, and the ignition takes place in a position further from the center. Our results shows that differences in the 20Ne central abundances remain modest, below 14%.

astro-ph.SR↗

arXiv2023

Persistent and occasional: searching for the variable population of the ZTF/4MOST sky using ZTF data release 11

We present a variability, color and morphology based classifier, designed to identify transients, persistently variable, and non-variable sources, from the Zwicky Transient Facility (ZTF) Data Release 11 (DR11) light curves of extended and point sources. The main motivation to develop this model was to identify active galactic nuclei (AGN) at different redshift ranges to be observed by the 4MOST ChANGES project. Still, it serves as a more general time-domain astronomy study. The model uses nine colors computed from CatWISE and PS1, a morphology score from PS1, and 61 single-band variability features computed from the ZTF DR11 g and r light curves. We trained two versions of the model, one for each ZTF band. We used a hierarchical local classifier per parent node approach, where each node was composed of a balanced random forest model. We adopted a 17-class taxonomy, including non-variable stars and galaxies, three transient classes, five classes of stochastic variables, and seven classes of periodic variables. The macro averaged precision, recall and F1-score are 0.61, 0.75, and 0.62 for the g-band model, and 0.60, 0.74, and 0.61, for the r-band model. When grouping the four AGN classes into one single class, its precision, recall, and F1-score are 1.00, 0.95, and 0.97, respectively, for both the g and r bands. We applied the model to all the sources in the ZTF/4MOST overlapping sky, avoiding ZTF fields covering the Galactic bulge, including 86,576,577 light curves in the g-band and 140,409,824 in the r-band. Only 0.73\% of the g-band light curves and 2.62\% of the r-band light curves were classified as stochastic, periodic, or transient with high probability ($P_{init}\geq0.9$). We found that, in general, more reliable results are obtained when using the g-band model. Using the latter, we identified 384,242 AGN candidates, 287,156 of which have $P_{init}\geq0.9$.

astro-ph.IM↗

arXiv2023

Stellar Variability in Ground-Based Photometric Surveys: An Overview

Time-resolved ground-based surveys in general, and photometric ones in particular, have played a crucial role in building up our knowledge of the properties, physical nature, and the very existence of the many different classes of variable stars and transient events that are currently known. Here I provide a brief overview of these developments, discussing some of the stumbling blocks that had to be overcome along the way, and others that may still hamper further progress in the area. A compilation of different types of past, present, and future surveys is also provided.

astro-ph.SR↗

arXiv2022

Three Ultra-short Period Contact Eclipsing Binary Systems Mined from Massive Astronomical Surveys

We present the photometric analysis of three ultra-short period total eclipsing binaries in contact configuration, CRTS_J172718.0+431624, OGLE-BLG-ECL-000104, and OGLE-BLG-ECL-000012, mined from massive astronomical surveys. Using the available archival light curves (LCs) from Vista Variables in the Via Lactea (VVV), Optical Gravitational Lensing Experiment (OGLE), Zwicky Transient Facility (ZTF) and Catalina Sky Survey (CSS) in different passbands and new multi-band photometric observations with the 2.3 m Aristarchos telescope at Helmos Observatory, their relative physical parameters were derived. We explored the parameter space by using PIKAIA genetic algorithm optimizer. The best photometric solution and error budget estimation was adopted for each system through MCMC sampling of the global optimum. The approximate absolute parameters were derived for each contact system adopting an empirical mass-luminosity relation. All three systems have a mass ratio lower than 0.5. The exchange between primary and secondary depths of CRTS_J172718.0+431624 during the years 2016-2022 may be due to spot activity. In addition, we present a detailed analysis of the first well characterized shortest period contact eclipsing binary with total eclipses known so far (OGLE-BLG-ECL-000104). Thanks to VVV and OGLE LCs, new distances were derived for OGLE-BLG-ECL-000104 and OGLE-BLG-ECL-000012 using empirical period-luminosity relations. The origin and evolutionary status of all three ultra-short period contact binaries are thoroughly discussed in the context of the detached-binary formation channel.

astro-ph.SR↗

arXiv2022

A segmented period-luminosity relation for nearby extragalactic $δ$ Scuti stars

The period-luminosity relations (PLR) of Milky Way $δ$ Scuti ($δ$ Sct) stars have been described to the present day by a linear relation. However, when studying extragalactic systems such as the Magellanic Clouds and several dwarf galaxies, we notice for the first time a non-linear behaviour in the PLR of $δ$ Sct stars. Using the largest sample of $\sim 3700$ extragalactic $δ$ Sct stars from data available in the literature $-$mainly based on OGLE and SuperMACHO survey in the Large Magellanic Cloud (LMC)$-$ we obtain that the best fit to the period-luminosity ($M_V$) plane is given by the following piecewise linear relation with a break at $\log{P} = -1.03 \pm 0.01$ (or $0.093 \pm 0.002$ d) for shorter periods (sp) and longer periods (lp) than the break-point: $$M_V^{sp} = -7.08 (\pm 0.25) \log{P} -5.74 (\pm 0.29) ;\hspace{5pt} \log{P} < -1.03$$ $$M_V^{lp} = M_V^{sp} + 4.38 (\pm 0.32) \cdot (\log{P} + 1.03 (\pm 0.01));\hspace{5pt} \log{P} \geq -1.03$$ Geometric or depth effects in the LMC, metallicity dependence, or different pulsation modes are discarded as possible causes of this segmented PLR seen in extragalactic $δ$ Sct stars. The origin of the segmented relation at $\sim 0.09$ days remains unexplained based on the current data.

astro-ph.SR↗

arXiv2022

Semi-Supervised Classification and Clustering Analysis for Variable Stars

The immense amount of time series data produced by astronomical surveys has called for the use of machine learning algorithms to discover and classify several million celestial sources. In the case of variable stars, supervised learning approaches have become commonplace. However, this needs a considerable collection of expert-labeled light curves to achieve adequate performance, which is costly to construct. To solve this problem, we introduce two approaches. First, a semi-supervised hierarchical method, which requires substantially less trained data than supervised methods. Second, a clustering analysis procedure that finds groups that may correspond to classes or sub-classes of variable stars. Both methods are primarily supported by dimensionality reduction of the data for visualization and to avoid the curse of dimensionality. We tested our methods with catalogs collected from OGLE, CSS, and Gaia surveys. The semi-supervised method reaches a performance of around 90\% for all of our three selected catalogs of variable stars using only $5\%$ of the data in the training. This method is suitable for classifying the main classes of variable stars when there is only a small amount of training data. Our clustering analysis confirms that most of the clusters found have a purity over 90\% with respect to classes and 80\% with respect to sub-classes, suggesting that this type of analysis can be used in large-scale variability surveys as an initial step to identify which classes or sub-classes of variable stars are present in the data and/or to build training sets, among many other possible applications.

astro-ph.SR↗

arXiv2022

Near-infrared observations of RR Lyrae and Type II Cepheid variables in the metal-rich bulge globular cluster NGC 6441

NGC 6441 is a bulge globular cluster with an unusual horizontal branch morphology and a rich population of RR Lyrae (RRL) and Type II Cepheid (T2C) variables that is unexpected for its relatively high metallicity. We present near-infrared (NIR, $JHK_s$) time-series observations of 42 RRL, 8 T2Cs, and 10 eclipsing binary candidate variables in NGC 6441. The multi-epoch observations were obtained using the FLAMINGOS-2 instrument on the 8-m Gemini South telescope. Multi-band data are used to investigate pulsation properties of RRL and T2C variables including their light curves, instability strip, period--amplitude diagrams, and period--luminosity and period--wesenheit relations (PLRs and PWRs) in $JHK_s$ filters. NIR pulsation properties of RRL variables are well fitted with theoretical models that have canonical helium content and mean-metallicity of NGC 6441 ([Fe/H]$=-0.44\pm0.07$~dex). The helium-enhanced RRL models predict brighter NIR magnitudes and bluer colors than the observations of RRL in the cluster. This suggests that these RRL variables in NGC 6441 are either not significantly helium enhanced as previously thought or the impact of such enhancement is smaller in NIR than the predictions of the pulsation models. We also use theoretical calibrations of RRL period--luminosity--metallicity (PLZ) relations to simultaneously estimate the mean reddening, $E(J-K_s)=0.26\pm0.06$~mag, and the distance, $d=12.67\pm0.09$~kpc, to NGC 6441. Our mean reddening value and the distance are consistent with the independent estimates from the bulge reddening map based on red clump stars and the latest Gaia data, respectively. Our distance and reddening values provide a very good agreement between the PLRs of T2Cs in NGC 6441 and those for RRL and T2Cs in Galactic globular clusters that span a broad range of metallicity.

astro-ph.SR↗

arXiv2022

Large amplitude periodic outbursts and long period variables in the VVV VIRAC2-$β$ database

The VISTA Variables in the Via Lactea (VVV) survey obtained near-infrared photometry toward the Galactic bulge and the southern disc plane for a decade (2010 - 2019). We designed a modified Lomb-Scargle method to search for large-amplitude ($Δ$Ks > 1.5 mag) mid to long-term periodic variables (P > 10 d) in the 2nd version of VVV Infrared Astrometric Catalogue (VIRAC2-$β$). In total, 1520 periodic sources were discovered, including 59 candidate periodic outbursting young stellar objects (YSOs), based on the unique morphology of the phase-folded light curves, proximity to Galactic HII regions and mid-infrared colours. Five sources are spectroscopically confirmed as accreting YSOs. Both fast-rise/slow-decay and slow-rise/fast-decay periodic outbursts were found, but fast-rise/slow-decay outbursts predominate at the highest amplitudes. The multi-wavelength colour variations are consistent with a variable mass accretion process, as opposed to variable extinction. The cycles are likely to be caused by dynamical perturbations from stellar or planetary companions within the circumstellar disc. An additional search for periodic variability amongst YSO candidates in published Spitzer-based catalogues yielded a further 71 candidate periodic accretors, mostly with lower amplitudes. These resemble cases of pulsed accretion but with unusually long periods and greater regularity. The majority of other long-period variables are pulsating dusty Miras with smooth and symmetric light curves. We find that some Miras have redder $W3 - W4$ colours than previously thought, most likely due to their surface chemical compositions.

astro-ph.SR↗

arXiv2022

The VVV survey: Long-period variable stars I. Photometric catalog of ten VVV/OGLE tiles

Long-period variable stars (LPVs) are pulsating red giants, primarily in the asymptotic giant branch phase, and they include both Miras and semi-regular variables (SRVs). Their period-age and period-luminosity relations enable us to trace different stellar populations, as they are intrinsically very bright and cover a wide range in distances and ages. The purpose of this study is to establish a census of LPV stars in a region close to the Galactic center, using the six-year database of the Vista Variables in the Vía Láctea (VVV) ESO Public Survey, as well as to describe the methodology that was employed to search for and characterize LPVs using VVV data. Near-IR surveys such as VVV provide a unique opportunity to probe the high-extinction innermost regions of the Milky Way. The detection and analysis of the intrinsically bright Miras in this region could provide us with an excellent probe of the properties of the Milky Way far behind its bulge. We used point-spread function photometry for all available $K_{s}$-band images in ten VVV tiles, covering $16.4~°^2$ in total, overlapping fields observed in the course of the Optical Gravitational Lensing Experiment (OGLE)-III survey. We designed a method to select LPV candidates, and we used the known variables from OGLE-III and other known variables from the literature to test our approach. The reduced $χ^2$ statistic, along with the flux-independent index $K_{(fi)}$, were used in our analysis. The Lomb-Scargle period search method, Fourier analysis, template fitting, and visual inspection were then performed to refine our sample and characterize the properties of the stars included in our catalog. A final sample of 130 Mira candidates, of which 129 are new discoveries, was thus obtained, with periods in the range between about 80 and 1400~days.

astro-ph.SR↗

arXiv2021

Period Change Rates in Large Magellanic Cloud Cepheids Revisited

The period-change rate (PCR) of pulsating variable stars is a useful probe of changes in their interior structure, and thus of their evolutionary stages. So far, the PCRs of Classical Cepheids in the Large Magellanic Cloud (LMC) have been explored in a limited sample of the total population of these variables. Here we use a template-based method to build observed minus computed (O-C) period diagrams, from which we can derive PCRs for these stars by taking advantage of the long time baseline afforded by the Digital Access to a Sky Century @ Harvard (DASCH) light curves, combined with additional data from the Optical Gravitational Lensing Experiment (OGLE), the MAssive Compact Halo Object (MACHO) project, Gaia's Data Release 2, and in some cases the All-Sky Automated Survey (ASAS). From an initial sample of 2315 sources, our method provides an unprecedented sample of 1303 LMC Classical Cepheids with accurate PCRs, the largest for any single galaxy, including the Milky Way. The derived PCRs are largely compatible with theoretically expected values, as computed by our team using the Modules for Experiments in Stellar Astrophysics (MESA) code, as well as with similar previous computations available in the literature. Additionally, five long-period (P>50 d) sources display a cyclic behavior in their O-C diagrams, which is clearly incompatible with evolutionary changes. Finally, on the basis of their large positive PCR values, two first-crossing Cepheid candidates are identified.

astro-ph.SR↗

arXiv2021

Using classical Cepheids to study the far side of the Milky Way disk. II. The spiral structure in the first and fourth Galactic quadrants

In an effort to improve our understanding of the spiral arm structure of the Milky Way, we use Classical Cepheids (CCs) to increase the number of young tracers on the far side of the Galactic disk with accurately determined distances. We use a sample of 30 CCs, discovered using near-infrared photometry from the VISTA Variables in the Vía Láctea survey (VVV) and classified based on their radial velocities and metallicities. We combine them with another 20 CCs from the literature for which VVV photometry is available. The compiled sample of CCs with homogeneously computed distances based on VVV infrared photometry was employed as a proof of concept to trace the spiral structure in the poorly explored far side of the disk. Although the use of CCs has some caveats, these variables are currently the only available young tracers in the far side disk for which a numerous sample with accurate distances can be obtained. Therefore, a larger sample could allow us to make a significant step forward in our understanding of the Milky Way disk as a whole. We present preliminary evidence that CCs favor: a spiral arm model with two main arms (Perseus and Scutum-Centaurus) branching out into four arms at galactocentric distances, $R_\mathrm {GC}\gtrsim5-6\,\mathrm{kpc}$; the extension of the Scutum-Centaurus arm behind the Galactic center; a possible connection between the Perseus arm and the Norma tangency direction. The current sample of CCs in the far side of the Galaxy are in the mid-plane, arguing against the presence of a severely warped disk at small Galactocentric distances ($R_\mathrm {GC}\lesssim12\,\mathrm{kpc}$) in the studied area. The discovery and characterization of CCs at near-IR wavelengths appears to be a promising tool to complement studies based on other spiral arm tracers and extend them to the far side of our Galaxy.

astro-ph.GA↗

arXiv2021

Informative Bayesian model selection for RR Lyrae star classifiers

Machine learning has achieved an important role in the automatic classification of variable stars, and several classifiers have been proposed over the last decade. These classifiers have achieved impressive performance in several astronomical catalogues. However, some scientific articles have also shown that the training data therein contain multiple sources of bias. Hence, the performance of those classifiers on objects not belonging to the training data is uncertain, potentially resulting in the selection of incorrect models. Besides, it gives rise to the deployment of misleading classifiers. An example of the latter is the creation of open-source labelled catalogues with biased predictions. In this paper, we develop a method based on an informative marginal likelihood to evaluate variable star classifiers. We collect deterministic rules that are based on physical descriptors of RR Lyrae stars, and then, to mitigate the biases, we introduce those rules into the marginal likelihood estimation. We perform experiments with a set of Bayesian Logistic Regressions, which are trained to classify RR Lyraes, and we found that our method outperforms traditional non-informative cross-validation strategies, even when penalized models are assessed. Our methodology provides a more rigorous alternative to assess machine learning models using astronomical knowledge. From this approach, applications to other classes of variable stars and algorithmic improvements can be developed.

astro-ph.IM↗