The fixed 53.647-day GJ 251 term improves joint infrared prediction under three tested noise models
We tested prediction at the two published radial velocity periods of GJ 251. We used all 624 rows in the published Figure 6 table. We omitted both infrared instruments, HPF and SPIRou, from training. A circular 53.647-day term improved their joint predictive log score by +7.65 nats under a shared activity grid. Independent instrument activity gave +9.63 nats. An exploratory test with separate instrument noise priors gave +10.06 nats. We integrated signal coefficients and finite priors for activity and extra noise. We removed constant instrument offsets. Separate omission tests gave positive scores for 4 of five instruments under the main grid. The small NEID and SPIRou scores changed sign with model or nearby period choices. We also tested coefficient priors and close observation times. These results apply to the tested models and processed velocities. Prior work selected the periods from the full data. This retrospective test does not give an independent planet discovery or a false-alarm probability. The package contains source tables, provenance, code, numerical checks, score arrays, and figures.
GJ 251 · radial velocity · predictive validation · Gaussian process · public data · instrument holdout
References
5- usesdoi:10.3847/1538-3881/ae0e20
We use the published Figure 6 velocity table, periods, effective wavelengths, and activity scales. Section 6.4.2 provides the prior instrument consistency test.
- usesdoi:10.5281/zenodo.17238981
We preserve two older example tables and the MIT notice to document their difference from the final published table.
- doi:10.1051/0004-6361/202038820
This work reports GJ 251 b and the rotation period near 122 days.
- doi:10.1051/0004-6361/202346472
This work describes the Wapiti processing used for the SPIRou velocities. We use processed velocities.
- An eccentric 647.6-day model lowers mean squared error in 9 of 18 omitted HD 20794 seasons in a fit with fixed periods and independent Gaussian noise sci:sha256:33d2a5aaa63141db39d6bccc13412d17b7337d9dca639abce667e5bdc293104c
This related study tests omitted-season prediction at published fixed periods for HD 20794. We compare its scope with our instrument omission test.
Cited by
4 publicationsPublic catalog entries on Sciscape.
- A blind Keplerian search of 743 HARPS-RVBank v2 stars with 30 or more nights recovers 248 of 465 known planets and finds 57 signals (FAP < 0.1 %) with no known match that pass activity tests
GJ 251 is not in our searched sample of 743 stars, so we give no result for this star.
- usesA fixed 37.90-day term improves Gl 725 B velocity prediction in three omitted years under three noise models with published correction vectors
This prior study informed our use of joint predictive log scores with fixed reported velocity periods. Our code and target data are separate.
- Six TESS sectors sample every GJ 251 c transit train with events of at least 1.94 hours in the tested White NEID box
This prior work tests the published GJ 251 radial velocity periods. The present work addresses observation-time coverage in TESS.
- usesTwo noise mixtures give a 0.29 to 1.54 nat gain for the 50.77-day term in 151 older GJ 887 velocities
We use its method of instrument prediction with orthonormal offset contrasts, proper Gaussian signal priors, and separate or shared activity. Our implementation and GJ 887 data are supplied here.