Introduction to astroML: Machine learning for astrophysics

Vanderplas, Jake; Connolly, Andrew J.; Ivezić, Željko; Gray, Alexander

doi:10.1109/cidu.2012.6382200

Cited by 192 publications

(126 citation statements)

References 26 publications

(33 reference statements)

Supporting

Mentioning

126

Contrasting

Order By: Relevance

“…The Lomb-Scargle routine in Numerical Recipes 2 , modified to return false alarm probability (FAP) values for all peaks, is valuable for the periodograms where there are comparatively clear-cut peaks, however for many of the spectroscopic results it is unable to return any clear periods. For these cases software was written in Python using alternative Lomb-Scargle routines provided by SCIPY library (Jones et al 2001), the ASTROML library, (Vanderplas et al 2012) and the GATSPY library, (VanderPlas & Ivezić 2015), comparing the results. We did this because in some cases the results widely differed and it gave an assessment of the stability of the calculations.…”

Section: Periodicity Of Proxima Centauri From Photometric Measurementsmentioning

confidence: 99%

Calculations of periodicity from Hαprofiles of Proxima Centauri

Collins

Jones

Barnes

2017

A&A

View full text Add to dashboard Cite

We investigate retrieval of the stellar rotation signal for Proxima Centauri. We make use of high-resolution spectra taken with UVES and HARPS of Proxima Centauri over a 13-yr period as well as photometric observations of Proxima Centauri from ASAS and HST. We measure the Hα equivalent width and Hα index, skewness and kurtosis and introduce a method that investigates the symmetry of the line, the peak ratio, which appears to return better results than the other measurements. Our investigations return a most significant period of 82.6 ± 0.1 days, confirming earlier photometric results and ruling out a more recent result of 116.6 days which we conclude to be an alias induced by the specific HARPS observation times. We conclude that whilst spectroscopic Hα measurements can be used for period recovery, in the case of Proxima Centauri the available photometric measurements are more reliable. We make 2D models of Proxima Centauri to generate simulated Hα, finding that reasonable distributions of plage and chromospheric features are able to reproduce the equivalent width variations in observed data and recover the rotation period, including after the addition of simulated noise and flares. However the 2D models used fail to generate the observed variety of line shapes measured by the peak ratio. We conclude that only 3D models which incorporate vertical motions in the chromosphere can achieve this.

show abstract

Section: Periodicity Of Proxima Centauri From Photometric Measurementsmentioning

confidence: 99%

Calculations of periodicity from Hαprofiles of Proxima Centauri

Collins

Jones

Barnes

2017

A&A

View full text Add to dashboard Cite

show abstract

“…to avoid saturation. We checked the light curves of all 55 of these objects for any signs of variability using a Lomb-Scargle periodogram via the astroML python package (Vanderplas et al 2012) and derived V magnitudes, based on the median value of those data, for objects with 3 or more epochs of data (see Fig. 3 for an example).…”

Section: Photometry From the Catalina Surveymentioning

confidence: 99%

Search for associations containing young stars (SACY)

Elliott

Bayo

Melo

et al. 2016

A&A

View full text Add to dashboard Cite

Context. The young associations offer us one of the best opportunities to study the properties of young stellar and substellar objects and to directly image planets thanks to their proximity (<200 pc) and age (≈5−150 Myr). However, many previous works have been limited to identifying the brighter, more active members (≈1 M ) owing to photometric survey sensitivities limiting the detections of lower mass objects. Aims. We search the field of view of 542 previously identified members of the young associations to identify wide or extremely wide (1000−100 000 au in physical separation) companions. Methods. We combined 2MASS near-infrared photometry (J, H, K) with proper motion values (from UCAC4, PPMXL, NOMAD) to identify companions in the field of view of known members. We collated further photometry and spectroscopy from the literature and conducted our own high-resolution spectroscopic observations for a subsample of candidate members. This complementary information allowed us to assess the efficiency of our method. Results. We identified 84 targets (45: 0.2−1.3 M , 17: 0.08−0.2 M , 22: <0.08 M ) in our analysis, ten of which have been identified from spectroscopic analysis in previous young association works. For 33 of these 84, we were able to further assess their membership using a variety of properties (X-ray emission, UV excess, H α , lithium and K I equivalent widths, radial velocities, and CaH indices). We derive a success rate of 76-88% for this technique based on the consistency of these properties. Conclusions. Once confirmed, the targets identified in this work would significantly improve our knowledge of the lower mass end of the young associations. Additionally, these targets would make an ideal new sample for the identification and study of planets around nearby young stars. Given the predicted substellar mass of the majority of these new candidate members and their proximity, high-contrast imaging techniques would facilitate the search for new low-mass planets.

show abstract

“…Because of the relatively large 1 Eyer & Bartholdi (1999) claimed that for most practical cases, lower periodicities (higher frequencies) can be detected even for strongly (but randomly) under-sampled observations. 2 Calculated by using the astroML python module (Vanderplas et al 2012) and its bootstrapping package. 3 http://exoplanets.org/code/ uncertainties in the RV and incomplete coverage of the RV curve, we also decided to fix the eccentricity of the orbit to the slightly non-circular value determined by Lillo-Box et al (2014b), e = 0.066 +0.013 −0.017 .…”

Section: Resultsmentioning

confidence: 99%

Radial velocity confirmation of Kepler-91 b

Lillo-Box

Barrado

Henning

et al. 2014

A&A

View full text Add to dashboard Cite

The object transiting the star Kepler-91 was recently assessed as being of planetary nature. The confirmation was achieved by analysing the light-curve modulations observed in the Kepler data. However, quasi-simultaneous studies claimed a self-luminous nature for this object, thus rejecting it as a planet. In this work, we apply an independent approach to confirm the planetary mass of Kepler-91b by using multi-epoch high-resolution spectroscopy obtained with the Calar Alto Fiber-fed Echelle spectrograph (CAFE). We obtain the physical and orbital parameters with the radial velocity technique. In particular, we derive a value of 1.09 ± 0.20 M Jup for the mass of Kepler-91b, in excellent agreement with our previous estimate that was based on the orbital brightness modulation.

show abstract

Introduction to astroML: Machine learning for astrophysics

Cited by 192 publications

References 26 publications

Calculations of periodicity from Hαprofiles of Proxima Centauri

Calculations of periodicity from Hαprofiles of Proxima Centauri

Search for associations containing young stars (SACY)

Radial velocity confirmation of Kepler-91 b

Contact Info

Product

Resources

About