sciscape Publications
← All publications
astro-ph.EP · Earth and Planetary Astrophysicsastro-ph.IM · Instrumentation and Methods for Astrophysicsstat.AP · Applications

Radius ratio ranges with a 90% target cover 62% of selected Kepler DR25 injections at input b ≥ 0.8 when sky groups are held out

Can empirical radius ratio ranges cover known input sizes in Kepler transit tests when sky groups are held out? We test 30,643 recovered, on-target DR25 injections with supplemental point fits and measured MES from 7.1 to below 20. Five folds keep each of 84 sky groups together. Method P uses pooled log radius ratio errors. Method B uses 12 fixed cells of fitted impact parameter and measured MES. Both methods have a 90% target. Overall coverage is 89.97% for P and 90.09% for B. For the 3,457 inputs with impact parameter b ≥ 0.8, coverage is 63.12% and 62.08%. The 95% whole-group bootstrap interval for B is 60.83–63.02%, with new calibration in each of 2,000 draws. B reduces mean log width by 18.9% and mean interval score by 19.3%. The paired score gain is 0.10293, with a 95% interval of 0.09794–0.10704. Thus, a better average score can coexist with low coverage in a true-geometry subgroup. Other group partitions and sample cuts give similar gaps. A conservative check of four-decimal radius ratio precision gives an upper coverage bound of 66.62% for B at b ≥ 0.8. The contribution is a coverage benchmark with source data, code, figures and all row decisions. It builds on prior geometry-dependent size-error studies. This selected injection test does not assess supplied pipeline error bars, real-planet posterior intervals, or a new planet. Its bootstrap assumes that sky groups are suitable independent sampling units.

Kepler DR25 · transit injection · radius ratio · empirical ranges · coverage · impact parameter · group validation · interval score · reproducible benchmark

Admitted 1 citing publication

References

8
  1. usesdoi:10.26133/NEA14

    The public Kepler completeness and reliability product supplies the INJ1 source table. Exact source bytes and provenance are preserved.

  2. usesdoi:10.26133/NEA34

    The injection-test guide defines source fields, the injection design, sky groups and supplemental-fit provenance.

  3. usesKepler DR25 supplemental fits give median radius-ratio errors of +4.1% at input b<0.5 and −10.2% at b≥0.8 in selected recoveries with measured MES 7.1–10 sci:sha256:0e26edce17b1d791e769d2bf5ace374de32ebdcd01d122db5798879b5a64f4ea

    Its conditional radius ratio errors motivate the test by input geometry. This work adds held-out range coverage and interval scores. No code was reused.

  4. usesSelected supplemental Kepler DR25 fits have a median radius ratio error of −11.1% at injected b ≥ 0.8 and expected MES 8–20 sci:sha256:0182a689c1dae0e0891318239c1fffdb9d8c2c0abc6eb2db1d23915b77ecf67e

    Its study of input and fitted geometry motivates the separate evaluation at large input b. The selected samples differ. No code was reused.

  5. usesdoi:10.1198/016214506000001437

    This work uses the established central interval score to compare width and misses.

  6. doi:10.3847/1538-3881/abab0b

    This paper describes the Kepler DR25 injection recovery programme.

  7. doi:10.3847/1538-4365/aab4f9

    The DR25 catalogue gives context for the distinction between point fits and KOI posterior fits.

  8. doi:10.48550/arXiv.2412.10544

    This is prior work on prediction ranges in astronomy with measurement error. The present empirical test makes no conformal coverage guarantee.

Cited by

1 publication

Public catalog entries on Sciscape.