The seven sisters DANCe

Abstract

Context. The photometric and astrometric measurements of the Pleiades DANCe DR2 survey provide an excellent test case for the benchmarking of statistical tools aiming at the disentanglement and characterisation of nearby young open cluster (NYOC) stellar populations. Aims. We aim to develop, test, and characterise of a new statistical tool (intelligent system) for the sifting and analysis of NYOC populations. Methods. Using a Bayesian formalism, with this statistical tool we were able to obtain the posterior distributions of parameters governing the cluster model. It also used hierarchical bayesian models to establish weakly informative priors, and incorporates the treatment of missing values and non-homogeneous (heteroscedastic) observational uncertainties. Results. From simulations, we estimated that this statistical tool renders kinematic (proper motion) and photometric (luminosity) distributions of the cluster population with a contamination rate of 5.8 ± 0.2%. The luminosity distributions and present day mass function agree with the ones found in a recent study, on the completeness interval of the survey. At the probability threshold of maximum accuracy, the classifier recovers ≈90% of the recently published candidate members and finds 10% of new ones. Conclusions. A new statistical tool for the analysis of NYOC is introduced, tested, and characterised. Its comprehensive modelling of the data properties allows it to get rid of the biases present in previous works. In particular, those resulting from the use of only completely observed (non-missing) data and the assumption of homoskedastic uncertainties. Also, its Bayesian framework allows it to properly propagate observational uncertainties into membership probabilities and cluster velocity and luminosity distributions. Our results are in a general agreement with those from the literature, although we provide the most up-to-date and extended list of candidate members of the Pleiades cluster.

Corona-Australis DANCe

“…We are currently measuring the proper motions of faint sources based on archival images, and our own observations, as part of the DANCe project (Bouy et al 2013) to complement the Gaia-DR2 catalogue in this region and to extend upon the methodology developed by Olivares et al (2018) in order to perform membership analysis in regions of high (and variable) extinction. In addition, we are also starting an observing cam-paign to characterise the newly discovered YSOs in this study.…”

“…Although our new sample of cluster members significantly improves the current census of stars in this region, we argue that our list is not complete yet. Our team is currently refurbishing the methodology developed by Olivares et al (2018), which was based on hierarchical Bayesian models, to perform membership analysis in regions of high extinction, and we will soon be able to provide a more complete census of the stars in the densest cores of Corona-Australis.…”

Ruprecht 147 DANCe

“…In the literature there are several methodologies that deal with the detection of cluster members (e.g. Olivares et al 2018a;Gagné et al 2018;Cantat-Gaudin et al 2018b;Sarro et al 2014). However, neither of them allow to extract candidate members from data sets of millions of sources while combining astrometric and photometric features and their uncertainties and correlations.…”

“…The methodology of Sarro et al (2014) is able to work on data sets of millions of sources taking into account both astrometric and photometric informations, but does not include the observational uncertainties in the construction of the cluster model. In our opinion, the methodology of Olivares et al (2018a) has clear advantages over the others: takes into account the full covariance matrix of the observations, incorporates photometry, and deals with partially observed objects. However it is so computationally expensive that it is impractical for our data set made of millions of sources.…”

The nature of the NGC 2546: Not one but two open clusters

“…The proper modeling of field star distribution required a large number of Gaussian components. This is not unexpected but a frequent result when all star data are included in a highly contaminated region (see for example Hosek et al 2015;Olivares et al 2018). The main reason is that the parameter distributions for the field are usually not Gaussian (Cabrera-Cano & Alfaro 1990;Sánchez & Alfaro 2009).…”