Table 6 Clusters of dimensions for machine-generated big data sources Cluster References Accuracy [38] Completeness, Significance [38] Consistency [38, 47] Trustworthiness [38] Freshness [38] Dimensions in italic are representative of the cluster.
Table 5 Clusters of dimensions for process-mediated big data sources Cluster References Accuracy, Reliability [12, 36, 37] Consistency [37] Redundancy [37] Spread, Value of the tail, Connectivity [10, 37] Copying [37] Freshness, Coverage [45] Dimensions in italic are representative of the cluster.
In contrast, subspace clustering would further associate each cluster with a subset of dimensions, such that each cluster would contain data that are only similar in its associated subspace S ⊂ ℜ N. Subspace clustering can be further classified as disjoint or overlapping.
Biclustering is the clustering of both dimensions in a single dataset (e.g. both genes and experiments in a gene expression dataset).
In addition, it has been found that the clustering of wellbeing dimensions is explained by one underlying common genetic effect (Bartels and Boomsma 2009).
On one hand, the understanding of compact clusters in subsets of dimensions is challenging itself.
In summary, we can see that even for results with 4 clusters, the majority of dimensions are from the given set above.
When taking a problem-oriented and trans-domain approach, a cluster of other integration dimensions are implicated.
These definitions imply the following two conditions: i) the set of data in a child cluster is a superset of the data in its parent clusters and ii) the set of dimensions in a child cluster is a subset of the data in its parent clusters.
A second point related to the subtractive clustering algorithm is that it requires to optimize the ratio of cluster dimensions for the best performance.
