Since the heuristic results tend to oversplit the taxa into too many species, it would be interesting to compare the (single lambda and multiple lambda) scores of the ground truth data with the scores obtained by the heuristic. This will show whether the heuristic is just too bad or the currently used Maximum-Likelihood model does not suit the given data.
Since the heuristic results tend to oversplit the taxa into too many species, it would be interesting to compare the (single lambda and multiple lambda) scores of the ground truth data with the scores obtained by the heuristic. This will show whether the heuristic is just too bad or the currently used Maximum-Likelihood model does not suit the given data.