I recall in 1980's, style-analyzing software reporting word counts for words with high frequency in the text but low frequency in general usage - jargon. The suggestion was to re-word, and it was often a relevant clue to re-think by examining why one word was doing so much work. Detecting that sounds like more than repetition, but still feasible probabilistically and relevant at both usage and conceptual levels.
deepsquirrelnet|9 months ago
w10-1|9 months ago
rsynnott|9 months ago