IJRCS – Volume 1 Issue 1 Paper 1


Author’s Name : Malarmathi M

Volume 01 Issue 01  Year 2014  ISSN No:  2349-3828  Page no: 1-7


Abstract -The knowledgesets that area unit within the sort of object-attribute-time format is remarked as three-dimensional (3D) data sets. clump these three-dimensional (3D) knowledge sets may be a troublesome task. that the mathematical space clump methodology is applied to cluster the three-dimensional (3D) knowledge sets. however finding the subspaces within the these three-dimensional (3D) dataset that is dynamic  over time is actually a troublesome task. typically this mathematical space clump on three-dimensional (3D) knowledge sets might manufacture the big range of arbitrary  and spurious clusters. thus to cluster these three-dimensional (3D) knowledge sets a brand new centre of mass based mostly thought is introduced referred to as CATS. This CATS permits the users to pick out the popular objects as centroids. This algorithmic rule isn’t the parallel one. thus it will increase the time and area needs that area unit required to cluster the three-dimensional (3D) knowledge sets. And in CATS no best centroids are chosen to cluster the three-dimensional (3D) datasets. Since the CATS clusters the information supported the mounted centroids, the CATS cannot manufacture the great quality clusters. thus for the primary time within the planned methodology the CPSO technique is introduced on the three-dimensional (3D) knowledge sets to beat of these drawbacks that clusters the three-dimensional (3D) datasets supported the best centroids and additionally it acts because the parallelization technique to tackle the area and time complexities.

Keywords –Component3D mathematical space clump, singular worth decomposition, numerical improvement technique, supermolecule structural knowledge analysis, money and stock knowledge analysis