I am trying to identify the Named entities from the text and categorize them into People, Places, and Organization. I am using Cloud Natural Language API by Google and Open Calais API to identify the named entities.
When I input a text containing the word 'China', Google NL API identify it as 'Person' type. However, its contextual meaning in the document addresses it as a Country. Is Google NL API able to give the entities based on the text context? If so, please let me know what I am missing.
If the text contains a word 'obama', google NL API outputs 'Obama' as Person whereas Open Calais API identifies 'Barak Obama' as a Person. Why is it so? What are the other ways to get the exact named entities out of the term in the text like the one which Open Calais returns?
With regard to (1), could you please share your sentence? I'm tried a few sentences like "I'm traveling to China" or "Do you know anyone in China", and it does return Location as the type for China (you can try quickly at https://cloud.google.com/natural-language/)
With regard to (2), NL API extracts the entity from the text. So, if the text has "Obama", it will extract "Obama", and if it's "barack obama", it extracts as so. However, it will resolve both cases correctly to the right wikipedia page and return the same MID for both cases. So, you can extract the information from those as well.