Routing Unreliable Risk Scores in Capacity Limited Decision Systems
The Eastasouth Journal of Information System and Computer Science · 30.12.2024 · 2 (02) · 386-397
DOI: 10.58812/esiscs.v2i02.1316 ↗Kurz gefasst
Fachartikel in der Zeitschrift The Eastasouth Journal of Information System and Computer Science, online erschienen am 30.12.2024. Verfasst von Saliheen Ahmaad Rayeef. Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).
Abstract (Originalsprache)
Organizations often have more flagged records than staff can review. A data-quality flag can trigger backup scoring or remove a record from the queue, but these responses have different consequences for event capture. This study compared four policies using 200 paired simulated cohorts of 4,000 records and a review budget of 600 per cohort. The primary score was more precise than the backup on clean records. A subset of primary scores was inflated, and an imperfect quality flag identified potentially corrupted scores. At 15% corruption, routing flagged records to backup scoring captured 217.9 events per cohort versus 171.5 with primary scoring alone. The mean paired difference was 46.3 events (95% Monte Carlo bootstrap interval [44.9, 47.8]). The clean scenario quantified the opportunity cost of using a less precise backup. These results characterize a queue allocation mechanism under specified assumptions and provide a reproducible design for testing the mechanism on operational data.
Zitieren
APA
Rayeef, S. A. (2024). Routing Unreliable Risk Scores in Capacity Limited Decision Systems. The Eastasouth Journal of Information System and Computer Science, 2(02), 386-397. https://doi.org/10.58812/esiscs.v2i02.1316
Vancouver
Rayeef SA. Routing Unreliable Risk Scores in Capacity Limited Decision Systems. The Eastasouth Journal of Information System and Computer Science. 2024;2(02):386-397. doi:10.58812/esiscs.v2i02.1316
BibTeX
@article{Rayeef2024,
author = {Rayeef, Saliheen Ahmaad},
title = {Routing Unreliable Risk Scores in Capacity Limited Decision Systems},
journal = {The Eastasouth Journal of Information System and Computer Science},
year = {2024},
volume = {2},
number = {02},
pages = {386--397},
doi = {10.58812/esiscs.v2i02.1316}
}
EUVIMED