A Note on Model Selection Based on the Percentage of Accuracy-Precision
Abstract
Mathematical methods such as empirical correlations, analytical models, numerical simulations, and data-intensive computing (data-driven models) are the key to the modeling of energy science and engineering. Accrediting of different models and deciding on the best method, however, is a serious challenge even for experts, as the application of models is not limited only to estimations, but to predictions and derivative properties. In this note, by combining meaningful metrics of accuracy and precision, a new metric for determining the best-in-class method was defined.