Accuracy/precision and coverage/recall cutpoints

[pingfreud]

Very interesting job to popularize Machine Learning :+1:

One thing bothered me though. The algorithm description page report Precision and Recall values, however no data is given about what cutpoint was used for the likelihood. It would be nice to report it.

[sacdallago]

@goldbergtatyana

[goldbergtatyana]

We used the machine learning algorithm at its default parameters, so the evaluation was done at the cut point of 50 (everything above = dead, everything below = alive).