There are certain challenges for the NER system, which are as follows:
- NER tools train on a closed domain dataset. So, an NER system developed for one domain does not typically perform well on an other domain. This requires a universal NER tool that can work for all domains, and after training it should able to generalize enough to deal with unseen situations.
- Sometimes you will find words which are the names of locations as well as the name of a person. The NER tool can't handle a case where one word can be expressed as the location name, person name, and organization name. This is a very challenging case for all NER tools. Suppose you have word TATA hospital; the single the words TATA can be the name of a person as well as the ...