Parametrisia

Parametrisia is a term used in some theoretical discussions to describe the study of parameters within mathematical, statistical, and computational models. It focuses on how parameter values, parameterization choices, and the geometry of the parameter space shape model behavior, prediction, and uncertainty.

Etymology: derived from parameter, with the -ia suffix commonly used to denote a field of study.

Overview: Parametrisia is not a formal science with a universal curriculum but a multidisciplinary perspective that sits at the intersection of statistics, systems theory, and applied modeling. Core concerns include identifying which parameters can be estimated from data (identifiability), how sensitive model outputs are to parameter changes (sensitivity analysis), and how to select, calibrate, and validate parameter values to reflect reality. It also covers methods for exploring parameter spaces, reducing models while preserving behavior, and quantifying residual uncertainty.

Methods: Tools from Bayesian inference and frequentist estimation are employed, including likelihood-based methods, prior modelling, calibration, and uncertainty quantification. Techniques such as profile likelihood, global optimization, and Sobol sensitivity indices are used to assess practical identifiability and to prioritize data collection.

Applications: Parametrisia informs modeling across domains such as pharmacokinetics, epidemiology, climate science, engineering, economics, and social sciences, where parameter choices critically affect forecasts and control strategies.

Status and reception: The concept is used variably and is not universally recognized as a formal discipline. It serves as a conceptual lens to emphasize parameter-driven aspects of modeling and decision making, alongside structural and theoretical considerations.

Related topics include parameter estimation, identifiability, sensitivity analysis, calibration, and model selection.