I have a problem when I want to detect person’s names. The names are very variable and it is impossible to have all possible names in the training data. In addition, people usually write only their name, for example if the name is requested by a form:
They only write:
“Lucas Frías” for example
So you don’t have context words like: My name is …, I am …
In this case it is more difficult to classify the message or recognize the name.
Instead of training your own model, have you tried using SpacyEntityExtractor for extracting people’s names? I’ve heard that it is good at extracting names.
Hi @saurabh-m523!, thanks for answering! It was really helpful!!
I tryed with Spacy and it seems that it works very well ! Only it’s too case sensitive and people are too creative when they write. So I am going to try use spacy and DIET to make it more robust.