|
KMID : 0379920020270010003
|
|
Journal of The Korea Socity of Health Informatics and Statistics
2002 Volume.27 No. 1 p.3 ~ p.12
|
|
|
Tree structured Clustering for Craniofacial Morphology
|
|
|
|
|
Abstract
|
|
|
Clustering is the task of segmenting a heterogeneous population into a number of more homogeneous subgroups or clusters. It is valuable method for understanding the complex nature of multivariate dental data and relationships. It is widely needed in growth pattern recongnition.
This paper tries to cluster with the 675 records together on the tree structured clustering. The data were collected from 675 records of 6-year old. From 675 lateral cephalograms, 136 variables from the X, Y coordinates of 32 landmarks were calculated. The landmarks vary with the growth pattern of craniofacial morphology. Clusters of 136 variables 32 land marks might indicate different pattern of growth.
The main aim is to use clustering techniques to develop an objective typology for classifying craniofacial morphology pattern. Clinically the purpose was that the derived clustering would be useful in identifying different pattern of the craniofacial growth which be of help in defining more directed orthodontic treatment plants and preventive orthodontics.
For the evaluation of a clustering procedure, we review the validity and the stability. The validity measured by seperation degree of each cluster. The variation of given statistics on the bootstrap sample indicates stable shapes of clusters. This variation can be interpreted as a measure of stability of clusters.
Finally we can get the diagrammatic cephalmetric representation for each cluster. 7 Craniofacial pattern was obtained with the highest validity and the best stability. There were significant differences in summarized variables among craniofacial patterns. The cranifacial patterns were characterized by the diagrammatic cephalometric representation.
|
|
|
KEYWORD
|
|
|
|
|
|
FullTexts / Linksout information
|
|
|
|
|
|
Listed journal information
|
|
|
|