Dear all,
is it possible to work with two InterminatedThresholds (say c1 and c2) instead of only one?
Here is a typical situation: we have a classification problem with two classes (0 and 1). We train a e.g. a neural network with a softmax layer.
After training, we want to we want to classify an object as "0" if the corresponding probability is larger than c1 and as "1" if the corresponding probability is larger than c2.
Currently, InterminateThreshold applies to all classes simulataneously.
Of course, it is easy to write some lines to solve this, but I wonder a bit if it is already implemented?
Can the UtilizationFunction be used somehow?