Page 1 of 1

Model validation

Posted: Thu Dec 03, 2020 3:48 am
by Dr John P Weldon
There is a model that applies to all instances of the system type modelled. Differences between individual instances are entirely attributable to input data (numbers).

In this situation what will constitute sufficient validation in terms of the number of individual instances validated? One instance; all instances; or some number in between?

Re: Model validation

Posted: Thu Dec 03, 2020 8:54 am
by Administrator
This question is a bit vague, it's hard to answer correctly.

General advice. Make sure the model passes the basic quality tests. Does it pass the units check? Do you have any mass-balance checks? Are all parameters clearly named and have a descriptive explanation? Are your sketches clear?