I have a lot of data and manually extracted annotations for the text. I was looking for any advice to automate the annotation extraction up to a good level of accuracy. Any advice is welcome.
There is a spectrum of methods. On one end is the tried n' true method of performing pattern matching with rules. Write a series of regular expressions (i.e., regex). On the other end is the newer methods of weak supervision, a system that learns over time to get better. Snorkel one of implementation of weak supervision. Here are some snorkel demos.