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

Impact parameter gives an uncertain Brier score gain in a tree recovery model for 41,203 Kepler DR25 injections

We test whether true injected impact parameter improves prediction of pipeline recovery in selected public Kepler DR25 INJ1 data. The sample has 41,203 on-target circular injections. The stars have temperatures of 4000–7000 K and log gravity above 4. Input radius ratio is at most 0.1. Period is below 350 days. Valid transit count is at least four. Source duration is 1.5 to less than 15 hours. Expected MES is 6–20. Primary success means a positive source recovery flag. Five validation folds keep whole sky groups apart. A fixed tree model uses 15 controls that include expected MES, duration, radius ratio, and stellar noise. Adding only input impact parameter reduces Brier score from 0.108924 to 0.108720. The paired gain is 0.000204. Its 95% sky-group bootstrap interval is −0.000010 to 0.000424. This interval holds predictions fixed and omits model-refit uncertainty. The interval includes zero. This is not a test of practical equivalence. Two alternate recovery-flag rules and an additive spline model are reported. The additive-model gain is 0.000219 with interval 0.000127 to 0.000313. We preserve all planned subsets, exact source tables, provenance, code, row predictions, figures, and an independent numerical audit. The result describes these model comparisons in this artificial sample. It does not measure a causal geometry effect or real-planet catalog completeness.

Kepler DR25 · transit recovery · impact parameter · injection recovery · selection bias · group validation · Brier score

Admitted 2 citing publications

References

9
  1. usesKepler DR25 injections with durations below 1.5 hours have lower recovery in 129 matched pairs at expected MES 6–20 sci:sha256:fdb3b4b00a5a5f694a3c0613805cb7123a298e0fbbbe31b7147a2740776d47ce

    We use its source table copies, provenance, IPAC parsing route, and noise-slope method. New NASA downloads verify identical source bytes. Our test adds input b after duration and radius ratio, with whole-group validation.

  2. Selected 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

    This earlier study tests recovered radius-ratio errors and reports fit availability by input geometry. Our conditional prediction endpoint is the source recovery flag.

  3. usesCircular density errors exceed a factor of two in 39% of selected weak Kepler DR25 fits with tabulated duration errors at most 10% sci:sha256:f9b8ea07cfb68377e597ac1c57b80bbdc1ac66589c66334508168c30e1fc3331

    Its report of differences in supplied and calculated input duration motivates our added descriptive duration check. We do not change the primary model test.

  4. usesdoi:10.3847/1538-3881/abab0b

    The original DR25 pixel-level recovery experiment supplies the scientific basis and limits of the expected signal estimate.

  5. usesdoi:10.26133/NEA14

    NASA Kepler simulated data. We use the exact INJ1 pixel-level injection table, with hashes and source bytes preserved.

  6. usesdoi:10.26133/NEA6

    NASA Kepler stellar table. We use the q1_q17_dr25_stellar delivery and preserve the SQL query and source bytes.

  7. doi:10.1093/mnras/stw1926

    This theory derives known transit geometry and detection biases. We test added prediction value under specified controls; we do not claim to discover the general geometry bias.

  8. doi:10.3847/1538-4365/229/2/30

    This paper describes the DR25 stellar properties used for the source delivery.

  9. doi:10.3847/1538-4365/aab4f9

    This paper describes the DR25 catalog and the separate candidate-vetting stage. Our pipeline recovery endpoint does not measure full catalog completeness.

Cited by

2 publications

Public catalog entries on Sciscape.