Extraction projects are used to create text intelligence models that extract information from the text of documents.
Within each project, users have to:
- Determine which classes of information need to be extracted.
- Collect a set of training documents and at least one set of test documents containing occurrences of all the possible information classes.
- Annotate the sets of documents with the expected results.
- Experiment with the creation of the model based on the annotated training set and apply the model to the test sets.
- Evaluate the results on the test sets.
- Until results are satisfactory, adjust the sets of documents and the annotations and repeat from step 4.
- Release and publish the model for practical use in NL Flow.
In this section of the manual you will find all the information you need to use Platform to perform all the above operations.