Models developed by mathematicians/statisticians based on criterion such as goodness of fit often leads to a “best” model only for the data utilized. Moreover the parameters in such models often do not have physical interpretations and as such their validity cannot be checked by other means. This article makes argument against modeling processes that do not incorporate information from discipline related to the origin of data and presents an example to demonstrate benefits of doing so.
Noubary, Reza D.
Mathematical Modeling, a Small Step in a Right Direction,
Applications and Applied Mathematics: An International Journal (AAM), Vol. 6,
1, Article 27.
Available at: https://digitalcommons.pvamu.edu/aam/vol6/iss1/27