J/A+A/708/A148 Estimating the peak energy of Swift GRBs (Sun+, 2026)
Estimating the peak energy of Swift gamma-ray bursts using supervised
machine learning.
Sun W.-P., Zhu S.-Y., Ma D.-L., Zhang F.-W.
<Astron. Astrophys. 708, A148 (2026)>
=2026A&A...708A.148S 2026A&A...708A.148S (SIMBAD/NED BibCode)
ADC_Keywords: Gamma rays ; GRB ; Models
Keywords: gamma-ray bursts: general - methods: statistical: machine learning
Abstract:
Gamma-ray bursts (GRBs) are among the most energetic explosive
phenomena in the universe, and their peak energy (Ep) is a key
physical quantity for understanding the prompt emission mechanism.
However, due to the limited energy coverage of the Swift satellite, a
large fraction of Swift GRBs lack reliable measurements of the peak
energy. Therefore, developing an accurate and efficient method to
predict Ep is of great importance. In this work, we propose a method
based on the SuperLearner framework that integrates multiple
supervised machine learning algorithms to predict Ep of Swift/BAT
GRBs. We use the Swift/BAT observational data from December 2004 to
September 2022 as training features, and adopt the peak energies of
516 GRBs jointly detected by Swift and either Fermi/GBM or Konus-Wind
as training labels. After training and testing multiple supervised
models, the final SuperLearner ensemble yields a more robust and
reliable predictive model. In 100 iterations of 5-fold cross
validation, the predicted E'p values show a tight correlation with the
observed Ep, with an average Pearson correlation coefficient of
r=0.72. Compared with previous Bayesian estimates, our model provides
predictions that are likely closer to the true values. Based on the
trained model, we further predict the peak energies of 650 Swift GRBs,
significantly increasing the number of GRBs with known peak energies
and providing new statistical support for constraining GRB emission
mechanisms and energy origins.
Description:
This dataset provides the observational parameters and predicted peak
energies for Swift/BAT gamma-ray bursts (GRBs). It includes three data
tables. table1.dat contains 516 BAT GRBs with reliable peak energy
measurements used as the training set. table3.dat presents the
generalization set of 650 BAT GRBs, providing their predicted and
bias-corrected peak energies estimated by the SuperLearner model.
table4.dat lists the properties of 392 BAT GRBs with measured
redshifts, including their rest-frame peak energies, isotropic
energies, and isotropic luminosities.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
table1.dat 65 516 List of 516 BAT GRBs in the training set
table3.dat 57 650 List of 650 BAT GRBs in the generalization set
table4.dat 119 392 Properties of BAT GRBs with redshift
--------------------------------------------------------------------------------
Byte-by-byte Description of file: table1.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 10 A10 --- GRB GRB name
12- 17 F6.2 s T90 Duration
19- 24 F6.2 ph/cm2/s Fp Peak flux
26- 30 F5.2 ph/cm2/s e_Fp ?=- 1-sigma symmetric error on Fp
32- 38 F7.2 10-14J/cm2 Sgamma Fluence (in 10-7erg/cm2)
40- 44 F5.2 10-14J/cm2 e_Sgamma 1-sigma symmetric error on Sgamma
(in 10-7erg/cm2)
46- 51 F6.2 --- Gamma Photon index
53- 56 I4 keV Ep Peak energy
58- 61 I4 keV E_Ep Upper error on Ep
63- 65 I3 keV e_Ep Lower error on Ep
--------------------------------------------------------------------------------
Byte-by-byte Description of file: table3.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 9 A9 --- GRB GRB name
11- 16 F6.2 s T90 Duration
18- 22 F5.2 ph/cm2/s Fp Peak flux
24- 27 F4.2 ph/cm2/s e_Fp ?=- 1-sigma symmetric error on Fp
29- 34 F6.2 10-14J/cm2 Sgamma Fluence (in 10-7erg/cm2)
36- 41 F6.2 10-14J/cm2 e_Sgamma ?=- 1-sigma symmetric error on Sgamma
(in 10-7erg/cm2)
43- 47 F5.2 --- Gamma Photon index
49- 52 I4 keV Eppred Predicted peak energy
54- 57 I4 keV Epcor Bias-corrected predicted peak energy
--------------------------------------------------------------------------------
Byte-by-byte Description of file: table4.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 10 A10 --- GRB GRB name
11 A1 --- n_GRB [*] Note on GRB name (1)
13- 18 F6.2 s T90 Duration
20- 25 F6.2 ph/cm2/s Fp Peak flux
27- 31 F5.2 ph/cm2/s e_Fp ?=- 1-sigma symmetric error on Fp
33- 39 F7.2 10-14J/cm2 Sgamma Fluence (in 10-7erg/cm2)
41- 45 F5.2 10-14J/cm2 e_Sgamma ?=- 1-sigma symmetric error on Sgamma
(in 10-7erg/cm2)
47- 51 F5.2 --- alpha Low-energy photon spectral index (2)
53- 58 F6.2 --- beta ? High-energy photon spectral index (2)
60- 66 F7.5 --- z Redshift
68- 71 I4 keV Epz ?=- Rest-frame peak energy
73- 77 I5 keV E_Epz ?=- Upper error on Epz
81- 84 I4 keV e_Epz ?=- Lower error on Epz
86- 90 F5.2 [10-7J] logEiso Isotropic energy (in erg)
92- 96 F5.2 [10-7J] E_logEiso Upper error on logEiso (in erg)
98-101 F4.2 [10-7J] e_logEiso Lower error on logEiso (in erg)
103-107 F5.2 [10-7W] logLiso Isotropic luminosity (in erg/s)
109-113 F5.2 [10-7W] E_logLiso Upper error on logLiso (in erg/s)
115-119 F5.2 [10-7W] e_logLiso Lower error on logLiso (in erg/s)
--------------------------------------------------------------------------------
Note (1): An asterisk (*) denotes GRBs in the generalization set.
Note (2): For BAT GRBs that can only be fitted with a simple power-law (PL)
model, we adopt alpha=-1 and beta=-2.25 of the Band function to calculate
the values of Eiso and Liso. If the best-fit spectral model is a cutoff
power-law (CPL), only the low-energy photon index (alpha) is listed
(beta is left blank).
--------------------------------------------------------------------------------
Acknowledgements:
Wan-Peng Sun, swp(at)glut.edu.cn
(End) Patricia Vannier [CDS] 05-Mar-2026