The ordering weight controls in what vertical order the rows are displayed in the row dendrogram. For column dendrograms it controls the horizontal order of the columns. The two subclusters within a cluster (there are always exactly two subclusters) are weighted and the cluster with the lower weight is placed above (to the left of) the other cluster. The weight can be any one of the following:
Input average rank of the rows (or columns). This is the order of the rows (or columns) during import to Spotfire.
Average value of the rows (or columns). For example, a row a with 5 dimensions would have the average (a1+a2+a3+a4+a5 )/5. The average for a row a with k dimensions is calculated as
Calculating the weight of a cluster
To calculate the weight w3 of a new cluster C3 formed from two subclusters C1 and C2 with a weight of w1and w2, and each containing n1 and n2 rows, you use the following expression:
See also: