I want to extract terminologies/multi-keywords (Unigram to five-gram) using machine learning. I have two questions:
What is the minimum number of labeled data needed
