Test the Model¶
Once each leaf category has 5 labels and you have fixed all the resulting conflicts by adding suggested features, select Test from the top-right navigation.
Download a test dataset and fill in according to the format provided, providing correct labels for each document. More information on test sets can be found here. Upload, click test and get the model prediction results. The predicted category appears selected at the top of a sorted list of categories by confidence score. Confidence scores designate the documents that are labeled correctly according to the test dataset. The one with the highest confidence score is highlighted in green.
If you want to edit your labelling decisions or the schema, press "Resume Training" in the top-right. Otherwise, press "Publish".
Lastly, you publish the model.