Information Technology Reference
In-Depth Information
Table 2.7
IFCM parameters
Parameters
Explanation
f
The input file name
c
The number of clusters, the default value is 3
m
The fuzzy factor, the default value is 2
w
The type of the sample weights, 0-equal (default), 1-user specified
s
The type of the initial centroids, 0-random (default), 1-user specified
i
The maximal number of iterations until convergence, the default value is 100
t
The threshold for stopping the iterations, the default value is 0.001
Table 2.8
Descriptions of two cluster validity criteria
Validity criteria
Functional description
Optimal cluster number
p
j
=
1
u
ij
i
=
1
c
1
p
Partition coefficient
V
PC
=
argmax
c
(
V
PC
,
U
,
c
)
i
=
1
c
j
=
1
u
ij
log
u
ij
p
1
p
Classification entropy
V
CE
=−
argmin
c
(
V
CE
,
U
,
c
)
where
p
is the number of samples in the data set, and
c
is the number of clusters
Step 2
Calculate the membership degrees and the centroids iteratively. First,
according to Eq. (
2.154
), we have
⎡
⎤
0
.
401 0
.
317 0
.
281
⎣
⎦
0
.
215 0
.
252 0
.
533
0
.
289 0
.
231 0
.
480
0
.
896 0
.
054 0
.
051
0
.
166 0
.
631 0
.
203
U
(
0
)
=
0
.
319 0
.
390 0
.
291
0
.
179 0
.
213 0
.
607
0
.
000 0
.
000 1
.
000
1
.
000 0
.
000 0
.
000
0
.
000 1
.
000 0
.
000
Step 3
According to Eq. (
2.155
), update the centroids as follows:
⎡
⎣
.
,
.
.
,
.
.
,
.
0
365
0
382
0
838
0
084
0
782
0
153
.
,
.
.
,
.
.
,
.
V
(
1
)
=
0
762
0
151
0
677
0
136
0
586
0
258
0
.
678
,
0
.
211
0
.
574
,
0
.
206
0
.
666
,
0
.
165
⎤
0
.
625
,
0
.
182
0
.
207
,
0
.
700
0
.
190
,
0
.
737
⎦
0
.
488
,
0
.
203
0
.
707
,
0
.
221
0
.
509
,
0
.
457
0
.
361
,
0
.
516
0
.
369
,
0
.
561
0
.
667
,
0
.
130
Search WWH ::
Custom Search