EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Comparative Evaluation of Machine Learning Models for Clinically Reliable Alpha-Thalassemia Carrier Screening

Sheetal Chopra, Dr Maninder Singh

Computer and Decision Making: An International Journal · 01.11.2026 · 3 (6S) · 1020-1033

DOI: 10.1922/comdem.v3i6s.18885 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Computer and Decision Making: An International Journal, online erschienen am 01.11.2026. Verfasst von Sheetal Chopra und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Alpha-thalassemia carrier screening increasingly relies on machine learning models trained on routine complete blood count and HPLC-derived parameters; however, most studies emphasise predictive accuracy without examining whether model outputs are clinically reliable. This study presents a comparative evaluation of representative machine learning techniques for alpha-thalassemia carrier identification, focusing on discrimination, sensitivity, specificity, calibration, robustness, threshold behaviour, and clinical utility. Multiple model families were assessed across CBC-only, HPLC-only, combined, and engineered feature settings, with additional analyses under missingness, measurement noise, cross-validation variability, and cost-sensitive screening conditions. The findings show that accuracy alone can misrepresent screening safety, as models with similar accuracy may differ substantially in balanced accuracy, false-negative risk, probability calibration, and threshold stability. Combined CBC-HPLC representations produced more dependable performance than isolated modalities. Overall, the study demonstrates that clinically useful thalassemia screening models must be evaluated through reliability-centred criteria rather than single-metric performance reporting.

Zitieren

APA

Chopra, S., & Singh, D. M. (2026). Comparative Evaluation of Machine Learning Models for Clinically Reliable Alpha-Thalassemia Carrier Screening. Computer and Decision Making: An International Journal, 3(6S), 1020-1033. https://doi.org/10.1922/comdem.v3i6s.18885

Vancouver

Chopra S, Singh DM. Comparative Evaluation of Machine Learning Models for Clinically Reliable Alpha-Thalassemia Carrier Screening. Computer and Decision Making: An International Journal. 2026;3(6S):1020-1033. doi:10.1922/comdem.v3i6s.18885

BibTeX
@article{Chopra2026,
  author = {Chopra, Sheetal and Singh, Dr Maninder},
  title = {Comparative Evaluation of Machine Learning Models for Clinically Reliable Alpha-Thalassemia Carrier Screening},
  journal = {Computer and Decision Making: An International Journal},
  year = {2026},
  volume = {3},
  number = {6S},
  pages = {1020--1033},
  doi = {10.1922/comdem.v3i6s.18885}
}
Kostenlos starten →