3 ms·
The problem is not the lack of unit tests: The problem is that for outside observers there are no tests at all that compare the model outputs to some ground dat
by VMG 6y ago
The problem is not the lack of unit tests: The problem is that for outside observers there are no tests at all that compare the model outputs to some ground data.
All the criticism would be moot if the model delivered sensible predictions.
- matthewdgreen 6y agoI think what you're saying is that you want to run the model on inputs from, say, February and see if it produces results that are similar to today's caseload.* So why don't you go do that? That isn't in any way a software engineering problem, it's a fundamentally difficult scientific one. * You'll obviously run into the "testing on your training data" problem, but again this isn't much of a criticism. We're in the midst of an epidemic and scientists are rapidly adjusting a model to reflect developing knowledge. The alternative is to have no models.