There is an automatically (intelligently) generated blog which I have read recently.
It appears to be (let’s give ‘seems’ some rest) quite a popular one in a certain section.
I know the corpus on which it was trained.
And the corpus on which it was retrained.
(Including most of the quotes and the comments, especially the long ones).
But I wonder whether the order of n-grams was five or six.
It is definitely better than four grams.
It could even be Se7en.
This brings up a new idea.
What about writing a paper on automatically guessing the order of n-grams, given some generated text?
It may be difficult in the general case, but in our case we know the corpus on which it was trained.
Any takers?
