(Today, 02:48 AM)npcompl33t wrote: You are not allowed to view links. Register or Login to view.I decomposed the vectors in the P/F vs K/T into actual word patterns.
A few patterns, K/T/P all move roughly the same on, such as 'od;
T and K mostly track each other, but diverge in specific areas.
P diverges from both K/T in other areas.
The most obvious are that Pcho transitions almost certainly into pche, a little lagged behind generic Te/Ke rises.
These graphs are very interesting, nice tool you built!
I have a general remark: at the moment we do not have a clear idea of which transitions should be studied. For instance, say 'chol' is progressively replaced by 'chedy'. It could be -ol becomes -edy, or 'che->cho' and 'l->dy', or 'ho->hed' and 'l->y' and so on. So the graphs could be informative if they focus on the correct transitions, but just show noise if they don't.
I have looked at some of your '21 rules' You are not allowed to view links. Register or Login to view., which effectively encode the path of transitions you have determined, I take the chance to ask you more about them. Maybe you explained it already and I missed it, but can you elaborate a little more on how they were found? How much are you confident in them?