On Examination of Medical Data with Approximate Reasoning - Fuzziness and Medicine - page 260

Table 18.1 Final cluster centers when traditional k -means clustering is used

Cluster
1 2 3 4 5 6 7
Age 57 56 57 47 64 62 63
Bmi 21 24 30 26 26 23 31
Syst 119 151 185 132 164 139 202

Table 18.2 Cluster centers when FCM is used

Cluster
1 2 3 4 5 6 7
Age 58 62 56 56 64 62 56
Bmi 30 26 24 24 23 31 22
Syst 184 165 139 151 140 201 120

In practice the FCM calculates the weighted distances between the data points
and the tentative cluster centers, and these weights base on the degrees of member-
ships of the points to the cluster centers. We obtain the optimal outcomes when the
centers in the densest clusters are selected.
Figure 18.3 depicts the degrees of membership of the persons to the first cluster
the center of which is Age 184. This means that one

typical group of persons in our data is having this approximate age, body mass index

= 58, Bmi = 30 and Syst =

Fig. 18.3 Degrees of membership of persons to the first cluster in Table 3.2 in our data set
when FCM is applied