A bayesian approach to robust modeling of skewed biomedical data

Creative Commons License

Cengiz M. A., Öztürk Z., Dünder E.

Scientific Reports, cilt.16, sa.1, 2026 (SCI-Expanded, Scopus)

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 16 Sayı: 1
  • Basım Tarihi: 2026
  • Doi Numarası: 10.1038/s41598-026-46263-w
  • Dergi Adı: Scientific Reports
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, BIOSIS, Chemical Abstracts Core, EMBASE, MEDLINE, Directory of Open Access Journals, Zoological Record, Academic Search Ultimate (EBSCO), Natural Science Collection (ProQuest), Biological Science Database (ProQuest), Biomedical Reference Collection: Corporate Edition (EBSCO), Health Research Premium Collection (ProQuest)
  • Anahtar Kelimeler: Bayesian inference, Biomedical datasets, Log-symmetric regression, Asymmetric data, MCMC
  • Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
  • Ondokuz Mayıs Üniversitesi Adresli: Evet

Özet

This article proposes a robust Bayesian approach to regression for modelling continuous, strictly positive, and asymmetric biomedical data. The methodology is based on log-symmetric distributions, offering flexibility in capturing skewness and accommodating heavy-tailed behaviour. Robustness is achieved through the careful specification of prior distributions and full posterior inference is conducted using Markov Chain Monte Carlo (MCMC) techniques. This allows for accurate estimation, comprehensive uncertainty quantification, and resilience to outliers. Model performance is evaluated through both classical and fully Bayesian selection criteria, including AIC, BIC, DIC, ICOMP, WAIC, and PSIS-LOO, providing a thorough assessment of fit and predictive ability. The effectiveness and adaptability of the proposed approach are demonstrated through extensive simulation studies and applications to real biomedical datasets involving physiological and diagnostic measurements. These results highlight the practical value of robust Bayesian regression for modern biomedical research, especially when distributional asymmetry and inferential reliability are of primary concern.