I have a question about multi-label modeling and metadata
I have a question about multi-label modeling.
I am analyzing data whose output is a graph value.
Usually, I first convert the label values into metadata and then specify the labels in the options of the multi-label modeling operator.
However, because I believed that the graph value data was related to each other, I proceeded with model learning without converting it to metadata.
The model training went well and the model was saved.
However, when I tried to put input into this model and see the predicted value, an error appeared saying that there was no label column.
In my opinion, I think it is strange to request label data from an already trained model just because label data is used for learning.
Did I do something wrong?