A Multistage Statistical Framework for Curve‐Family Model Selection

Empirical models are widely used to represent families of experimental curves in chemical engineering phenomena, such as drying kinetics, thermal degradation kinetics, osmotic dehydration, and equilibrium moisture isotherms; however, the absence of coherent procedure exists for identifying the best performing models. Selection relies on the visual inspection of a small set of tabulated heterogeneous indicators, a practice especially limited in drying kinetics due to the large number of available models. An analysis of 52 studies published between 1999 and 2023 revealed persistent subjectivity in model‐selection practices and a marked underuse of available statistical tools. On the other hand, simulation‐based validation provided robust evidence of the consistency and reliability of this framework.