What's the right way of checking whether or not the text is in a specific language?

[patrickschur]

What's the right way of checking whether or not the text is in a specific language?

[patrickschur]

The library computes the similarity of the N-grams of your input text with the N-grams of the language files. So if your input text has the exact same N-grams (and the order is also the same) as in one of the language files, you should get a score of 1. If none of the N-grams is found in one of the language files, you should get a 0.

If you get a negative score, you probably forgot to train the library and the number of maxNgrams is not the same as the number used during training. If this is not the case then something is wrong in the implementation.

[marcellofuschi]

Thank you so much for the clear explanation!