It is bunch of heuristics... just slightly more of them producing (in my opinion) slightly better result. The core of the article however points at the main pr

It is bunch of heuristics... just slightly more of them producing (in my opinion) slightly better result. The core of the article however points at the main problem: no evaluation dataset on which we could compare these algorithms.