Data-Based Mechanistic Modelling, Forecasting and Control

This book builds on the methods introduced in the author's previous Springer book Recursive Estimation and Time-Series Analysis to deliver a powerful and practical framework for Data-Based Mechanistic (DBM) modelling directly from time-series data. DBM modelling produces interpretable continuous-time transfer function models, allowing the underlying linear or nonlinear differential equations to be understood in clear physical terms closely tied to the real dynamics behind the data. All modelling tools are freely available in dedicated toolboxes for MATLAB, including advanced modules for forecasting and control based on DBM models. The four application chapters provide an in-depth perspective on DBM modelling, while a separate tutorial appendix offers a guided, step-by-step introduction to the complete DBM workflow using accessible hydrological examples. These methodological foundations are then illustrated by their application in three major areas of topical importance: global climate dynamics, the COVID-19 epidemic, and investment-unemployment interactions in the USA. This book also shows how large-scale simulation models can be distilled into compact, transparent DBM representations, and how these small 'emulation' models can deliver physical insight, reliable forecasting, and effective control (management) strategies. Blending rigorous estimation theory, real-world relevance, and fully reproducible tools, this book offers a unique bridge between advanced data-based modelling and today's most pressing dynamical problems.