Guttemberg data

[emanjavacas]

I've been thinking about the data and although it might sound a bit silly I just thought that if the generators fail to capture stylistic features at all (which might happen due to the reduced datasets per author that we have), we could always try the same thing on bigger datasets (book level) such as Guttemberg? I am just dropping this here for the record.

[emanjavacas]

Kind of irrelevant... I was going too pessimistic I suppose