Hi,
Is there any option of fit function that have at least one discrete parameter an some non-discrete parameter?
Hi,
there is only categories, but they are not necessarily parameters. You can use categories to split the data in different regions, and fit in each of them.
There is also a technical reason why discrete parameters do not work well: A gradient-descent minimiser needs gradients. However, calculating the derivative of a discrete parameter is a problem.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.