Forum: open-discussion
Monitor Forum | 
Start New Thread
| New idea [ Reply ] By: Rupert Collins on 2011-06-01 04:37 | [forum:4522] |
|
Was chatting to Stephane today, and we discussed a way of measuring how much discriminatory information was in each of our windows. The "mean distance of each window" measure is crude, as a few long branches could be driving the difference. What I think we need is a measure of information that is scaled. So, using the bootstrap might be an answer. We could measure bootstrap support for all clades (or just terminal clades), and then plot distribution of bootstrap support. Once you get to 100%, it cannot go further, so the measures should be more relative. |
|
