Abstract
A fundamental challenge for wide-field imaging surveys is obtaining follow-up spectroscopic observations: there are >109 photometrically cataloged sources, yet modern spectroscopic surveys are limited to ∼few× 106 targets. As we approach the Large Synoptic Survey Telescope era, new algorithmic solutions are required to cope with the data deluge. Here we report the development of a machine-learning framework capable of inferring fundamental stellar parameters (T eff, log g, and [Fe/H]) using photometric-brightness variations and color alone. A training set is constructed from a systematic spectroscopic survey of variables with Hectospec/Multi-Mirror Telescope. In sum, the training set includes ∼9000 spectra, for which stellar parameters are measured using the SEGUE Stellar Parameters Pipeline (SSPP). We employed the random forest algorithm to perform a non-parametric regression that predicts T eff, log g, and [Fe/H] from photometric time-domain observations. Our final optimized model produces a cross-validated rms error (RMSE) of 165 K, 0.39 dex, and 0.33 dex for T eff, log g, and [Fe/H], respectively. Examining the subset of sources for which the SSPP measurements are most reliable, the RMSE reduces to 125 K, 0.37 dex, and 0.27 dex, respectively, comparable to what is achievable via low-resolution spectroscopy. For variable stars this represents a ≈12%-20% improvement in RMSE relative to models trained with single-epoch photometric colors. As an application of our method, we estimate stellar parameters for ∼54,000 known variables. We argue that this method may convert photometric time-domain surveys into pseudo-spectrographic engines, enabling the construction of extremely detailed maps of the Milky Way, its structure, and history.
| Original language | English (US) |
|---|---|
| Article number | 122 |
| Journal | Astrophysical Journal |
| Volume | 798 |
| Issue number | 2 |
| DOIs | |
| State | Published - Jan 10 2015 |
Fingerprint
Keywords
- methods: data analysis
- methods: statistical
- stars: general
- stars: statistics
- stars: variables: general
- surveys
ASJC Scopus subject areas
- Space and Planetary Science
- Astronomy and Astrophysics
Cite this
A machine-learning method to infer fundamental stellar parameters from photometric light curves. / Miller, A. A.; Bloom, J. S.; Richards, J. W.; Lee, Y. S.; Starr, D. L.; Butler, Nathaniel; Tokarz, S.; Smith, N.; Eisner, J. A.
In: Astrophysical Journal, Vol. 798, No. 2, 122, 10.01.2015.Research output: Contribution to journal › Article
}
TY - JOUR
T1 - A machine-learning method to infer fundamental stellar parameters from photometric light curves
AU - Miller, A. A.
AU - Bloom, J. S.
AU - Richards, J. W.
AU - Lee, Y. S.
AU - Starr, D. L.
AU - Butler, Nathaniel
AU - Tokarz, S.
AU - Smith, N.
AU - Eisner, J. A.
PY - 2015/1/10
Y1 - 2015/1/10
N2 - A fundamental challenge for wide-field imaging surveys is obtaining follow-up spectroscopic observations: there are >109 photometrically cataloged sources, yet modern spectroscopic surveys are limited to ∼few× 106 targets. As we approach the Large Synoptic Survey Telescope era, new algorithmic solutions are required to cope with the data deluge. Here we report the development of a machine-learning framework capable of inferring fundamental stellar parameters (T eff, log g, and [Fe/H]) using photometric-brightness variations and color alone. A training set is constructed from a systematic spectroscopic survey of variables with Hectospec/Multi-Mirror Telescope. In sum, the training set includes ∼9000 spectra, for which stellar parameters are measured using the SEGUE Stellar Parameters Pipeline (SSPP). We employed the random forest algorithm to perform a non-parametric regression that predicts T eff, log g, and [Fe/H] from photometric time-domain observations. Our final optimized model produces a cross-validated rms error (RMSE) of 165 K, 0.39 dex, and 0.33 dex for T eff, log g, and [Fe/H], respectively. Examining the subset of sources for which the SSPP measurements are most reliable, the RMSE reduces to 125 K, 0.37 dex, and 0.27 dex, respectively, comparable to what is achievable via low-resolution spectroscopy. For variable stars this represents a ≈12%-20% improvement in RMSE relative to models trained with single-epoch photometric colors. As an application of our method, we estimate stellar parameters for ∼54,000 known variables. We argue that this method may convert photometric time-domain surveys into pseudo-spectrographic engines, enabling the construction of extremely detailed maps of the Milky Way, its structure, and history.
AB - A fundamental challenge for wide-field imaging surveys is obtaining follow-up spectroscopic observations: there are >109 photometrically cataloged sources, yet modern spectroscopic surveys are limited to ∼few× 106 targets. As we approach the Large Synoptic Survey Telescope era, new algorithmic solutions are required to cope with the data deluge. Here we report the development of a machine-learning framework capable of inferring fundamental stellar parameters (T eff, log g, and [Fe/H]) using photometric-brightness variations and color alone. A training set is constructed from a systematic spectroscopic survey of variables with Hectospec/Multi-Mirror Telescope. In sum, the training set includes ∼9000 spectra, for which stellar parameters are measured using the SEGUE Stellar Parameters Pipeline (SSPP). We employed the random forest algorithm to perform a non-parametric regression that predicts T eff, log g, and [Fe/H] from photometric time-domain observations. Our final optimized model produces a cross-validated rms error (RMSE) of 165 K, 0.39 dex, and 0.33 dex for T eff, log g, and [Fe/H], respectively. Examining the subset of sources for which the SSPP measurements are most reliable, the RMSE reduces to 125 K, 0.37 dex, and 0.27 dex, respectively, comparable to what is achievable via low-resolution spectroscopy. For variable stars this represents a ≈12%-20% improvement in RMSE relative to models trained with single-epoch photometric colors. As an application of our method, we estimate stellar parameters for ∼54,000 known variables. We argue that this method may convert photometric time-domain surveys into pseudo-spectrographic engines, enabling the construction of extremely detailed maps of the Milky Way, its structure, and history.
KW - methods: data analysis
KW - methods: statistical
KW - stars: general
KW - stars: statistics
KW - stars: variables: general
KW - surveys
UR - http://www.scopus.com/inward/record.url?scp=84924238350&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84924238350&partnerID=8YFLogxK
U2 - 10.1088/0004-637X/798/2/122
DO - 10.1088/0004-637X/798/2/122
M3 - Article
AN - SCOPUS:84924238350
VL - 798
JO - Astrophysical Journal
JF - Astrophysical Journal
SN - 0004-637X
IS - 2
M1 - 122
ER -