kyoniichan
commented
5 years ago one thing we did in class is use nltk and machine learning to make it judge if a name is male or female based on character length vs final letter final letter of the name was more accurate though both are imperfect it had like, 65% accuracy with final letter in determining gender of name i think it'd be a lot of work but we'd need like a corpus of human interactions like a bunch of variations on hello and appropriate responses etc., idk see https://www.nltk.org/book/ch06.html
Adding more AI to Diaglogflow feels like it would need more work. ATM, AI is very limited on greetings and such. Should consider removing it altogether?
If greetings is only the main use for this, maybe it would be better to parse it in the code instead.