A comprehensive review and unified taxonomy of intelligent hybrid fractional-order models incorporated with machine learning in healthcare applications
An International Journal of Optimization and Control: Theories & Applications (IJOCTA) · 20.11.2026 · 026370263
DOI: 10.36922/ijocta026370263 ↗Kurz gefasst
Fachartikel in der Zeitschrift An International Journal of Optimization and Control: Theories & Applications (IJOCTA), online erschienen am 20.11.2026. Verfasst von David Amilo und 4 weiteren Autor:innen.
Abstract (Originalsprache)
Fractional-order differential equations (FDEs) describe memory and hereditary effects that integer-order models cannot capture, while machine-learning (ML) methods extract patterns from large and noisy clinical data but often lack mechanistic meaning. Combining the two has become an active research direction in healthcare, yet the literature is spread across many disease areas and there is no common way to describe how the coupling is actually implemented. This review addresses that gap. Following the PRISMA 2020 guidelines, we searched PubMed and OpenAlex for journal articles published between January 2018 and March 2026 and used Scopus, Web of Science, and Google Scholar for citation searching. After screening 209 records, 43 hybrid FDE–ML studies were included. We propose a two-level taxonomy in which three coupling types (ML-informed estimation of fractional parameters; fractional physics-informed neural networks (FPINNs) and neural solvers; and fractional preprocessing of features) describe the mathematical link, and a fourth type describes the deployment layer (simulation engine and clinician-facing interface). We define a hybrid operator Q with an explicit domain and range that expresses every reviewed model as a composition of these components, and we illustrate it on the public UCI Dermatology dataset using repeated stratified cross-validation, baseline classifiers, and bootstrap confidence intervals. The evidence tables show that FPINNs and neural solvers dominate infectious-disease modelling, whereas oncology and cardiometabolic studies more often combine several types. Reported gains are study-specific, so we present them as representative values rather than pooled effects. We also discuss computational cost, data scarcity, uncertainty quantification, regulatory validation, and fairness. The main contribution of the work is a reproducible classification and a formal operator that can guide model selection, reporting, and validation in future hybrid fractional–ML studies in healthcare.
Zitieren
APA
Amilo, D., Farman, M., Kamran, M., Ravichandran, C., & Nisar, K. S. (2026). A comprehensive review and unified taxonomy of intelligent hybrid fractional-order models incorporated with machine learning in healthcare applications. An International Journal of Optimization and Control: Theories & Applications (IJOCTA), 0(0), 026370263. https://doi.org/10.36922/ijocta026370263
Vancouver
Amilo D, Farman M, Kamran M, Ravichandran C, Nisar KS. A comprehensive review and unified taxonomy of intelligent hybrid fractional-order models incorporated with machine learning in healthcare applications. An International Journal of Optimization and Control: Theories & Applications (IJOCTA). 2026;0(0):026370263. doi:10.36922/ijocta026370263
BibTeX
@article{Amilo2026,
author = {Amilo, David and Farman, Muhammad and Kamran, Muhammad and Ravichandran, C. and Nisar, Kottakkaran Sooppy},
title = {A comprehensive review and unified taxonomy of intelligent hybrid fractional-order models incorporated with machine learning in healthcare applications},
journal = {An International Journal of Optimization and Control: Theories & Applications (IJOCTA)},
year = {2026},
volume = {0},
number = {0},
pages = {026370263},
doi = {10.36922/ijocta026370263}
}
EUVIMED