search

matsen / pplacer

Phylogenetic placement and downstream analysis
http://matsen.fredhutch.org/pplacer/
GNU General Public License v3.0
74 stars 18 forks source link

try doing only approximate PAM #269

Closed matsen closed 12 years ago

matsen commented 12 years ago

Suggestion from reviewer:

  • In all the results, the PAM heuristic appears to have been run to convergence. However, iterative algorithms like PAM might approach the optimal solution very quickly then undergo a large number of iterations with only minor improvements before finally converging. For large problem sizes, e.g. 100k or more leaves, it would be very interesting to know if a nearly optimal solution can be achieved quickly by PAM. Related to this, is the improvement in ADCL by PAM at each iteration monotonically non-increasing?
habnabit commented 12 years ago

semi-punted