|
|
Re: Correct way to normalize an rmsd-based distance metric used in
repeated trials of pairs
Posted:
Mar 24, 2012 12:28 PM
|
|
Ray -
This is a simple question just to make sure I understand w = x -
mean[x] correctly.
For overall correlations in which we've plugged "zeroes" according to
your "input+1" and "output+0.5" suggestion, there are always 55 values
of x:
0, 0.5, 1, 1.5, . . . , 8, 8.5, 9
So if I use all 55 of these values to compute the mean, we get mean[x]
= 4.5
But of course, values duplicate in 0, 0.5, 1, 1.5, . . . , 8, 8.5, 9,
e.g. 2 from (1,3) and 2 from (2,2), and if I take the mean of the non-
redundant values, then w is again 4.5
So, I assume that I use mean[x] = 4.5 for the overall data, correct?
And by analogy:
mean[x] = 2.25 for the row (0,0),...,(0,9)
mean[x] = 2.75 for the row (1,0),...,(1,9)
mean[x] = 3.25 for the row (2,0),...,(2,9)
Correct?
If I'm incorrect regarding any of the above, please take a moment to
clarify.
Thanks very much again.
|
|
