Quantity one user guide – Bio-Rad Quantity One 1-D Analysis Software User Manual
Page 225
Quantity One User Guide
10-16
method of computing the new distance matrix in step 5. In the discussion that follows,
let:
p, q be indices indicating two clusters that are to be joined into a single cluster.
k be the index of the cluster formed by joining clusters p and q.
i be the index of any remaining clusters other than cluster p, q, or k.
n
p
the number of samples in the p’th cluster.
n
q
the number of clusters in the q’th cluster.
n the number of clusters in the k’th cluster formed by joining the p’th and q’th cluster.
n = n
p
+ n
q
d
pq
the distance between cluster p and cluster q.
Single Linkage
This is also called Nearest Neighbor or Minimum Method.
Complete Linkage
This is also called the Furthest Neighbor or Maximum Method.
Single and Complete linkage are good algorithms for indicating outlier clusters.
UPGAMA
Unweighted pair group method using arithmetic averages. This is also called
Weighted Average Linkage.
d
ki
min d
pi
d
qi
(
,
)
=
d
ki
max d
pi
d
qi
(
,
)
=
d
k i
n
p
n
-----
d
pi
⋅
n
q
n
-----
d
qi
⋅
+
=