Part of my core day job work involves identifying real vs not real people and also uses the same technology to drive everything - lucene. An algorithm that refuses to attempt identification beyond what data yelp itself stores is clearly in denial of the "mountain" of data available on people in the form of public web apis on the internet. I could write an "algorithm"/analyzer that does a better job than what yelp is currently doing in a day.
At the end of the day the biggest loser in all of this is probably yelp's users. Inaccurate reviews means you're getting inaccurate results on finding places local to you...which means the tool isn't nearly as useful to you as an end consumer. They might want to just sit down and find another way to get the same sales figures without sacrificing the quality of their product. Otherwise all it will take is someone else to come along and give people the right product to put them out of business. These stories that keep appearing on the internet about Yelp are increasingly becoming harder to ignore for everyone, which is sad for all the hard working engineers at Yelp who's only desire is to deliver a kick ass product. Lame