There is a growing body of literature examining the clustering of health risk behaviours, but little consensus about which risk factors can be expected to cluster for which sub groups of people.
The process of defining attributes involves examining the clusters of elements associated with the concept.
Examining the clustering of factors that relate to CVD, such as the metabolic syndrome and its components may warrant attention.
In examining the clusters for general trends, it was found that for most of the clusters, regulation on D-glucose and starch are very similar.
Examining the clustering patterns of multiple modifiable health risk factors and demographics associated with health clusters allows targeting of health prevention interventions towards co-occurring risk factors and for the subgroups most likely to exhibit these risk clusters [ 22, 50].
GENECLUSTER looks for evidence of potential additional association signals by examining the clustering of case and control haplotypes at the tips of a genealogical tree estimated from reference-panel data (here HapMap CEU) at a fine grid of locations.
We examined the cluster plots for 1008 SNPs and found that 36 of these were poor; these resided in the proximity of 35 of the significant SNPs.
We also examined the cluster analysis for each medical school separately, followed by a split half cross-validation procedure for each medical school.
Two studies examined the cluster stability of the empirically derived subgroup classification based on the Multidimensional Pain Inventory and found that retest resulted in 65 to 70% of patients being assigned to the same cluster [ 29, 30].
The study also examined the clustering of manufacturing companies in particular regions.
