Mathematical and Statistical Modeling of Uncertainty: A Comprehensive Review of Theoretical Foundations and Applied Methods

Abstract

Uncertainty affects scientific research, engineering design, economic forecasting, healthcare decision-making, climate modeling, artificial intelligence, and policy analysis, but its mathematical and statistical treatment remains fragmented across disciplines. This review provides a comprehensive, non-systematic synthesis of theoretical foundations and applied methods for modeling, quantifying, propagating, and interpreting uncertainty. Foundational books, seminal studies, and recent peer-reviewed literature were reviewed conceptually and thematically. The article examines uncertainty classifications, mathematical frameworks, statistical techniques, propagation methods, complex-system modeling, applied domains, comparative evaluation, challenges, and future directions. The review shows that uncertainty arises from randomness, incomplete knowledge, model assumptions, parameter estimation, measurement error, and computational limitations. Probability theory, fuzzy sets, interval analysis, evidence theory, and possibility theory provide complementary representations, while Bayesian inference, Monte Carlo simulation, bootstrap resampling, and sensitivity analysis support quantification and propagation. Applications extend across engineering reliability, finance, climate modeling, healthcare, artificial intelligence, and supply chain systems. Future research should emphasise hybrid, interpretable, scalable, and application-sensitive frameworks that integrate probabilistic, non-probabilistic, computational, and data-driven approaches for reliable decision-making under complex uncertainty.