Hacker News Reader

>> items w/ low coverage but high precision

Newbie NLP dev here. What's this concept of coverage vs precision that you speak of? It sounds intriguing.

It's the ratio of false positives and false negatives to their true equivalents.