I'm using their list for a project that incorporates some techniques of authorship analysis. I don't know what to do with "it's", though. I obviously cannot ignore the apostrophe and count it as "its", nor should I simply tokenise the components because then I would be over-counting the ratio of "it" to "is".
The question is, then, do I count "'s" as equivalent to "is", or do I leave out all occurrences of "it's" altogether? Of course, I know I could do either, but what would be the better choice in terms of preserving statistical distinctions amongst authors? At the moment I'm inclined to count "'s" as equivalent to "is", but I don't have good reasons for it.