I used this for a project re: similarity between two strings. The Jaccard similarity between sets of uni- and bi-grams was a surprising effective metric. DOG

I used this for a project re: similarity between two strings. The Jaccard similarity between sets of uni- and bi-grams was a surprising effective metric. DOG -> {d, o, g, do, og} GOD -> {g, o, d, go, od} intersection = {d, g, o} union = {d, g, o, do, go, od, og} J = 3 / 7 = ~43%