Comment by costco

3 years ago

I could have used a part of speech tagger, looked at time of day a user posts, capitalization, spelling errors, etc. From what I understand the state of the art is lightyears ahead of this, there are even companies with actual linguists who will act as expert witnesses in court to say stuff like "we can say with 95% certainty that xyz authored this email." Honestly it's kind of scary. There are papers that talk about cross platform authorship attribution, one I think did it with Twitter, Blogspot, G+ and had pretty good results.