Closed simongray closed 8 years ago
Perhaps I should start with just implementing a good regexner: http://nlp.stanford.edu/software/regexner.html
This can be used for more complex stuff, e.g. "I am a
In case I need to overwrite standard results: http://stackoverflow.com/questions/32247485/is-it-possible-to-train-a-stanford-ner-model-with-a-data-set-containing-regular
Usage example for TokensRegex: http://stackoverflow.com/questions/31747631/how-to-chain-rules-using-tokensregex
The semantic tagging from issue #23 deprecates this idea.
Similar to #13, this is to be a way for the user to decide what way s/he wants to discover entities.
One possible issue is that by using the default NER (and gender) annotator, valuable resources have already been spent performing that sort of tagging. Perhaps it can be removed from the pipeline and initiated independently in a separate class?