J/AJ/159/41 Data for ∼550 exoplanets using a neural network (Tasker+, 2020)
Estimating planetary mass with deep learning.
Tasker E.J., Laneuville M., Guttenberg N.
<Astron. J., 159, 41 (2020)>
=2020AJ....159...41T 2020AJ....159...41T
ADC_Keywords: Exoplanets; Stars, masses
Keywords: Computational methods ; Astronomy databases ; Exoplanet catalogs ;
Neural networks ; Astrostatistics tools
Abstract:
While thousands of exoplanets have been confirmed, the known
properties about individual discoveries remain sparse and depend on
detection technique. To utilize more than a small section of the
exoplanet data set, tools need to be developed to estimate missing
values based on the known measurements. Here, we demonstrate the use
of a neural network that models the density of planets in a space of
six properties that is then used to impute a probability distribution
for missing values. Our results focus on planetary mass, which neither
the radial velocity nor transit techniques for planet identification
can provide alone. The neural network can impute mass across the four
orders of magnitude in the exoplanet archive, and return a
distribution of masses for each planet that can inform us about trends
in the underlying data set. The average error on this mass estimate
from a radial velocity detection is a factor of 1.5 of the observed
value, and 2.7 for a transit observation. The mass of Proxima Centauri b
found by this method is 1.6-0.36+0.46M⊕, where the upper and
lower bounds are derived from the root mean square deviation from the
log mass probability distribution. The network can similarly impute
the other potentially missing properties, and we use this to predict
planet radius for radial velocity measurements, with an average error
of a factor 1.4 of the observed value. The ability of neural networks
to search for patterns in multidimensional data means that such
techniques have the potential to greatly expand the use of the
exoplanet catalog.
Description:
The neural network employed is a modified Boltzmann machine generative
model (Ackley+ 1985, Cognitive Science 9 147; Chen & Murray 2003, IEE
Proceedings-Vision, Image and Signal Processing 150 153). The network
learns the joint distribution of available exoplanet properties in
order to generate new data points that lie within the same
distribution. The properties chosen comprise six observables: planet
mass, planet radius, orbital period, stellar mass, equilibrium
temperature, and the number of known planets in the system. Our data
was taken from the NASA exoplanet archive. To ensure the largest
possible data set, missing values of our six planet properties in the
confirmed planet record on the exoplanet archive were compared with
the associated Kepler Object of Interest entry where available, and
that value used if present.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
fig1.dat 72 550 Distributions of radii, planetary masses, numbers
of planets in the system, stellar masses,
orbital periods and equilibrium temperatures
in the training and test data (see Section 2)
--------------------------------------------------------------------------------
See also:
B/corot : CoRoT observation log (N2-4.4) (CoRoT, 2016)
V/133 : Kepler Input Catalog (Kepler Mission Team, 2009)
IV/34 : K2 Ecliptic Plane Input Catalog (EPIC) (Huber+, 2017)
J/ApJ/622/1102 : The planet-metallicity correlation. (Fischer+, 2005)
J/ApJ/715/1050 : Predicted abundances for extrasolar planets. I. (Bond+, 2010)
J/ApJ/733/68 : Exoplanet masses derived from RVs (Brown+, 2011)
J/ApJ/761/123 : KELT-1 photometry and spectroscopy follow-up (Siverd+, 2012)
J/ApJ/767/95 : Stellar parameters of smallest KIC stars (Dressing+, 2013)
J/ApJ/794/159 : Statistical analysis of exoplanet surveys (Brandt+, 2014)
J/A+A/563/A143 : WASP-68b, WASP-73b, WASP-88b transits (Delrez+, 2014)
J/ApJ/787/80 : 139 Kepler planets transit time variations (Hadden+, 2014)
J/ApJS/210/25 : Transit timing variation for planetary pairs. II. (Xie, 2014)
J/A+A/594/A63 : International Deep Planet Survey results (Galicher+, 2016)
J/AJ/152/204 : HARPS-N radial velocities of HD 179070 (Lopez-Morales+, 2016)
J/ApJ/825/19 : Mass-rad. relationship for planets with Rp<4 (Wolfgang+, 2016)
J/A+A/602/A107 : 231 transiting planets eccentricity and mass (Bonomo+, 2017)
J/ApJ/834/17 : Mass & radius of planets, moons, low mass stars (Chen+, 2017)
J/AJ/153/93 : MOST photometry of Proxima (Kipping+, 2017)
J/A+A/606/A107 : K2/HARPS measurements for 8 stars (Oshagh+, 2017)
J/AJ/153/136 : Planets & their host stars with Gaia plx (Stassun+, 2017)
J/AJ/156/264 : California-Kepler Survey. VII. Planet rad. gap (Fulton+, 2018)
J/AJ/157/235 : Observations of the Kepler field with TESS (Christ+, 2019)
J/A+A/630/A135 : Beyond the exoplanet mass-radius relation (Ulmer-Moll+, 2019)
http://exoplanetarchive.ipac.caltech.edu/ : NASA exoplanets archive
Byte-by-byte Description of file: fig1.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 27 A27 --- ID Planet identifier
29- 33 F5.3 Rjup Rad [0.03/3.6] Planet radius
35- 40 F6.3 Mjup Mass [0/28] Planet mass
42- 55 F14.8 d Per [0.24/60191] Planet orbital period
57- 63 F7.2 K Teq [50/7719] Equilibrium temperature (1)
65 I1 --- N [1/8] Number of known planets in the system
67- 72 F6.3 Msun M* [0.08/24] Stellar mass
--------------------------------------------------------------------------------
Note (1): The equilibrium temperature was calculated from the stellar radius
(R*), stellar effective temperature (T*), and average orbital distance
(<αp>) via Teq=T*(R*/(2.0<αp>))^0.5 where
no value was entered in the exoplanet catalog.
--------------------------------------------------------------------------------
History:
From electronic version of the journal
(End) Prepared by [AAS], Coralie Fix [CDS], 18-Mar-2020