This involves grouping together words into chunks based on their part of speech, usually using regular expressions.