Java Data Mining Concepts - Java Data Mining: Strategy, Standard, and Practice - page 164

60

30

40

20

20

10

0 0

Values

Values

(a) Histogram for AGE (b) Histogram for INCOME

Rule:
IF
AGE <= 31.5 and AGE >= 15.0 and INCOME <= 67500.0 and INCOME >= 5000.0
THEN
Cluster equal 1
Support = 6/10=0.6

Figure 7-13 Cluster 1 details.

the attribute histograms and the rule of Cluster 1. Users can visually
inspect the attribute value distributions in each cluster to understand
how they differ. In this case Cluster 1 can be described as customers
whose age is between 15 and 31.5 and whose income is between $5,000 and
$67,500.

7.5.5 Apply Clustering Model: Assign New Cases to the Clusters

A clustering model can be applied to the original cases or new cases
to assign each case a cluster ID. Clustering apply also computes the
probability that the case belongs to the cluster, and the distance of the
case from the centroid of the cluster to assess how well each case fits
a given cluster. Like classification, JDM supports assigning the top
cluster, mapping the top or bottom clusters for each case, and per
cluster probabilities.