what is the proximity matrix between two clusters AB and CD given the distance matrix
Answers
I have a (symmetric) arrange M that addresses the partition between each pair of center points. For example,
A B C D E F G H I J K L
A 0 20 40 60 100 120
B 20 0 20 60 80 120 140
C 20 0 20 60 80 120 140
D 20 0 60 80 120 140
E 40 60 0 20 60 80
F 60 80 20 0 20 40 60
G 60 80 20 0 20 60 80
H 60 80 20 0 60 80
I 100 120 60 40 60 0 20
J 120 140 80 60 80 20 0 20
K 120 140 80 60 80 20 0 20
L 120 140 80 60 80 20 0
Is there any system to expel bunches from M (if fundamental, the amount of packs can be fixed), to such a degree, that each gathering contains center points with little detachments between them. In the point of reference, the gatherings would be (A, B, C, D), (E, F, G, H) and (I, J, K, L).
I've adequately endeavored UPGMA and k-infers yet the resulting bunches are astoundingly horrible.
The divisions are the ordinary stages a sporadic walker would take to go from center point A to center point B (!= An) and come back to center A. It's guaranteed that M^1/2 is an estimation. To run k-suggests, I don't use the centroid. I describe the detachment between center n bunch c as the typical partition among n and all center points in c.