I recently came across an article that attempts to use machine learning to identify Satoshi Nakamoto. The author collected hundreds of documents from seven potential candidates for the title of the real Satoshi Nakamoto and analyzed them using document classification techniques. I haven't seen this approach before and was intrigued by the results which pointed to Gavin Andresen as the most plausible candidate based on his word choice. An obvious criticism of this is that of all the authors, Andresen was in contact with Nakamoto and may have picked up some of his word usage. Additionally, he and Nakamoto discussed Bitcoin explicitly and extensively, whereas writings by some of the other suspects never mention it because they predate Bitcoin. Writing in a post-Bitcoin world may have introduced other terms (e.g. blockchain) into Andresen's vocabulary which could link his writing style more closely to Nakamoto's. Regardless, I think it's an interesting approach to the modern Satoshi Nakamoto mystery and is worth exploring.
