Beyond Point Estimates: Bayesian Deep Nonparametric Regression with Rigorous Uncertainty Quantification
Jurnal Pendidikan Matematika · 26.08.2025 · 2 (4) · 13
DOI: 10.47134/ppm.v2i4.2052 ↗Kurz gefasst
Fachartikel in der Zeitschrift Jurnal Pendidikan Matematika, online erschienen am 26.08.2025. Verfasst von Hasan Mohammed Iskander. Die Arbeit nennt 15 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Uncertainty quantification is essential in regression tasks where predictions inform high-stakes decisions. We present a practical framework for Bayesian deep nonparametric regression that moves beyond point estimates to deliver calibrated predictive intervals and uncertainty decomposition. The approach employs a heteroscedastic Bayesian neural network trained via Monte Carlo Dropout, enabling the estimation of both epistemic and aleatoric uncertainties without costly Markov chain Monte Carlo sampling. We evaluate the method on a synthetic heteroscedastic regression problem, demonstrating accurate predictive means, well-calibrated 90% prediction intervals, and computational efficiency on CPU-only hardware. The results highlight the method’s suitability for uncertainty-aware regression in resource-constrained settings, and all code is released for reproducibility.
Zitieren
APA
Iskander, H. M. (2025). Beyond Point Estimates: Bayesian Deep Nonparametric Regression with Rigorous Uncertainty Quantification. Jurnal Pendidikan Matematika, 2(4), 13. https://doi.org/10.47134/ppm.v2i4.2052
Vancouver
Iskander HM. Beyond Point Estimates: Bayesian Deep Nonparametric Regression with Rigorous Uncertainty Quantification. Jurnal Pendidikan Matematika. 2025;2(4):13. doi:10.47134/ppm.v2i4.2052
BibTeX
@article{Iskander2025,
author = {Iskander, Hasan Mohammed},
title = {Beyond Point Estimates: Bayesian Deep Nonparametric Regression with Rigorous Uncertainty Quantification},
journal = {Jurnal Pendidikan Matematika},
year = {2025},
volume = {2},
number = {4},
pages = {13},
doi = {10.47134/ppm.v2i4.2052}
}
EUVIMED