Catalog 1 bandwidth had an AUROC of 0.53 for later repeat labels in 473 CHIME sources
Original burst bandwidth gave little observed gain when ranking later repeat labels. We joined original CHIME/FRB Catalog 1 measurements to Catalog 2. The test used one original burst for each of 473 apparent single burst sources. Eleven had a later catalog repeat label. Narrower bandwidth gave an area under the receiver operating characteristic curve (AUROC) of 0.528. The 95% source bootstrap interval was 0.346–0.719. The narrowest 119 sources contained three positive labels. Their label fraction was 2.52%, compared with 2.33% in the full cohort. Removing seven uncertain sources left ten confirmed associations among 466 sources. Their AUROC was 0.557, with an interval of 0.366–0.754. Checks of quality flags, burst components, and source names gave similar estimates. Checks of exposure, signal-to-noise ratio, and dispersion measure gave no clear rank advantage. The observed gain for this rule was small. The intervals remain wide and include useful positive discrimination. Repeat detection depends on exposure and sensitivity. This result does not measure intrinsic repetition or test models with several features. The package supplies a source-level benchmark, exact data provenance, code, figures, results, and reproduction steps.
fast radio bursts · CHIME · repeating sources · spectral bandwidth · source-level evaluation · catalog cross-match · repeat detection · selection effects · reproducible astronomy
References
11- usesdoi:10.11570/21.0007
Original Catalog 1 release 210806.vFINAL supplies the burst measurements and original source status.
- doi:10.3847/1538-4365/ac33ab
The first CHIME/FRB catalog paper describes the original sample and measurements.
- usesdoi:10.11570/25.0066
Catalog 2 release 20251231 supplies later source associations, burst dates, and observing exposure.
- doi:10.3847/1538-4365/ae3828
The second CHIME/FRB catalog paper describes the later observations and inclusion rules.
- usesdoi:10.11570/23.0004
The 2023 source JSON identifies the 25 confirmed repeaters used in the source status check.
- doi:10.3847/1538-4357/acc6c1
The 2023 repeater paper gives the confirmed and candidate source groups.
- usesdoi:10.48550/arXiv.2605.08410
Version 1, Table 1, Table 2, and Table A supply updated confirmed and candidate source status facts.
- doi:10.3847/1538-4357/ac33ac
Prior work found narrower spectra among bursts from known repeating sources. This result motivates the tested direction.
- doi:10.3847/1538-4357/adb623
Prior classification work already tested fourteen sources whose repeat status changed after Catalog 1. The present work tests a fixed bandwidth rule with a full unlabelled comparison group.
- doi:10.3847/1538-4357/ae3c85
Prior work used spectral features to estimate repetition probability with Catalog 1 and the 2023 repeater release.
- doi:10.1007/s11433-026-3036-5
Prior work studied repeater features with Catalog 2. Its method is not evaluated in this fixed-rule test.
Cited by
2 publicationsPublic catalog entries on Sciscape.
- Earlier pulse data gave uncertain prediction gains for 19 repeating CHIME/FRB sources
This related study predicts later repeat labels from Catalog 1 bandwidth. Our target is later measured pulse shape within sources with several accepted days. No code or measurements were reused.
- Only one of 19 joint 2022 FRB candidates has a repeat label in CHIME Catalog 2
This study tests a bandwidth rule in the same catalog cohort and reports source-name and source-identity checks. We test the fixed published candidate lists.