More details about achieving a good quality ratio

[sgarciafer]

First of all thank's for the great work and sharing it on GitHub.
I've installed and played around with it, first using the sample language files, then adding a couple of books by language to the learn system.

Though the success ratio is quite poor... It would be nice to have some explanations and use cases to know how much and what kind of text should be given to the system to learn and achive a good ratio of success.