1995
DOI: 10.1109/91.413225
|View full text |Cite
|
Sign up to set email alerts
|

On cluster validity for the fuzzy c-means model

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

5
637
0
21

Year Published

2002
2002
2021
2021

Publication Types

Select...
5
4

Relationship

0
9

Authors

Journals

citations
Cited by 1,600 publications
(663 citation statements)
references
References 9 publications
5
637
0
21
Order By: Relevance
“…É uma prática comum em abordagens sistemáticas, para a identificação da quantidade ideal de grupos em problemas de agrupamento automático, utilizar k=2..k max , sendo k max = n ( [Pal and Bezdek, 1995] [Pakhira et. al., 2005] .…”
Section: -Experimentos Computacionaisunclassified
“…É uma prática comum em abordagens sistemáticas, para a identificação da quantidade ideal de grupos em problemas de agrupamento automático, utilizar k=2..k max , sendo k max = n ( [Pal and Bezdek, 1995] [Pakhira et. al., 2005] .…”
Section: -Experimentos Computacionaisunclassified
“…In detail: The two levels (DE and FSTS method) run iteratively until a minimum BIC value is achieved. We propose the user to run OLYMPUS after setting the cluster number in the range {N cmin , N cmax }, where N cmin is set to 2 and N cmax equal to g n according to the rule of thumb used by many investigators in literature [30][31]. The behavior of BIC function monotonically decreases until it hits the minimum at the correct cluster number and monotonically increases as the c n increases (see Supplementary File 2 -Fig.1).…”
Section: Bayesian Information Criterionmentioning
confidence: 99%
“…It is worth nothing that at the previous step (nearest neighbor search) we define ffiffiffi n p clusters, so in order to take advantage of previous calculations, the number of representative samples employed in our process is ffiffiffi n p . Since this is considered an upper bound for the number of groups of instances in a dataset [34], it is expected that most groups have a representative among the samples.…”
Section: Sampling and Projectionmentioning
confidence: 99%