Among them, the use of sampling techniques has proved to be especially useful for data clustering methods that perform several iterations considering different initializations, such as CLARANS [19] and PAM [2].
Several methods for data selection, clustering and functional analysis are available; however, these steps are usually performed independently, without exploiting and integrating the information derived from each step of the analysis.
For data clustering, the clustering stability methods on model selection techniques have been investigated in [14].
In "Related work", existing methods for data stream clustering which includes primitive and latest study with respect to their weaknesses and strengths will be reviewed.
The Self-Organizing Map (SOM) [4] is one of the best known unsupervised methods for data visualization and clustering [5, 6].
The NJ algorithm is free from 'molecular clock' restrictions and is widely accepted as a standard method for distance data clustering in biology (41).
Consequently, hierarchical modelling is strongly advocated as a more appropriate statistical method for dealing with data clustered within hospitals [ 38- 42].
For the adjusted analysis we will perform regression analysis of individual level data using methods for clustered data, multilevel, hierarchical or random effects models [ 37, 43].
Since the unit of randomisation is the PHCC, we will use a regression analysis of individual level data using methods for clustered data [ 38].
The small sample size and number of events in some participating centres is a limitation in our analysis because we modelled the mortality using methods for clustered data with centres as clusters.
