Supervised method to automatically learn the threshold

[hardbyte]

Consider adding a train function that would be provided with training data - CLKs that are known to match. The output would be an optimal threshold t.

This would be a vast improvement over the current method of essentially guessing a similarity threshold.

Ideally considering issue #23 - this should be able to work on a vector of clks.

Aha! Link: https://csiro.aha.io/features/ANONLINK-77