I created a model for the OpenNLP NE tagger node. When using it as an external model in the node, I still have to choose a preset tag labelling (such as: person, location, etc.) to name the found entities. Is there a possibility to make an own labelling?
e.g.: I created an model to tag diseases. Now I want the node to use the model and tag relevant tokens as 'disease'.
As a workaround I currently use 'person' random label. So if it's not possible to use own labels, maybe someone got an idea on how to rename the tag afterwards?
I am not a Text Mining expert, but I tried to achieve what you want.
Unfortunately I think this is not possible. The tags in the OpenNLP NE tagger aren't simple strings but represent complete models. Even when you use your own dictionary the library tries to load the model.
But this is just my impression, maybe someone else knows more about the OpenNLP stuff.
this is not possible with this node. You need to specify a tag that the node will use in the dialog. In the next major version 3.3 we will release a model learner node (based on the Stanford nlp lib) that can be trained on other tag sets beside the standard NE tag set. But until then you need to use your workaround.